import cx_Oracle as oracle
import pandas as pd
import matplotlib.pyplot as plt
from matplotlib.dates import AutoDateLocator, DateFormatter 
import numpy as np
import math
import seaborn as sns
import scipy.stats as stats  
import scipy
import scipy.cluster.vq
import scipy.spatial.distance
EuclDist = scipy.spatial.distance.euclidean
from mpl_toolkits.mplot3d import Axes3D
from factor_analyzer import FactorAnalyzer
import importlib
from sklearn import metrics,datasets,manifold
import copy
from sklearn import tree
import graphviz
import pydotplus
from six import StringIO
import os
from IPython.display import Image
import joblib
from sklearn.metrics import accuracy_score
from sklearn.metrics import precision_score
from sklearn.metrics import recall_score
from sklearn.model_selection import GridSearchCV

Get the data, and do the dimension reduction

Here, i get the data from the database, then do the dimension reduction by the method of factor analysis

Get data firstly.

df_feat = pd.read_csv("TotalSampleAfterProcessing.csv" , encoding = 'gbk')
df_feat.drop(columns=['Unnamed: 0'] , inplace = True)
df_feat.head(5)

	CAR_NUM	F1	F2	O1	D1	O2	D2	FS	STD_FD	AVG_FD
0	浙AR911W	0.809524	1.000000	0.857143	0.809524	0.666667	0.714286	0.630876	3.113023	6.090909
1	浙AJ183G	0.047619	0.380952	0.095238	0.142857	0.047619	0.047619	-1.105313	0.547723	1.600000
2	浙A916KS	0.285714	1.000000	0.190476	0.142857	0.380952	0.333333	-1.348978	2.097060	3.210526
3	浙A918CJ	0.285714	0.619048	0.142857	0.142857	0.238095	0.142857	-0.232944	5.589105	5.285714
4	浙AKB603	0.333333	0.571429	0.428571	0.380952	0.285714	0.285714	-0.341521	1.788854	3.200000

# 从dataframe中进行1%的抽样
df_feat_sample = df_feat.sample(n=None, frac=0.01, replace=False, weights=None, random_state=0, axis=None)
df_feat_sample = df_feat_sample.reset_index(drop=True)

#see the statistical indicators of the sample data
df_feat_sample.describe()

	F1	F2	O1	D1	O2	D2	FS	STD_FD	AVG_FD
count	4623.000000	4623.000000	4623.000000	4623.000000	4623.000000	4623.000000	4623.000000	4623.000000	4623.000000
mean	0.270037	0.606100	0.257635	0.239712	0.217608	0.229010	-1.008043	2.549642	4.127539
std	0.256971	0.286932	0.243612	0.236694	0.210898	0.212095	1.084510	1.936698	2.326158
min	0.047619	0.047619	0.047619	0.047619	0.047619	0.047619	-4.295169	0.000000	0.000000
25%	0.047619	0.380952	0.047619	0.047619	0.047619	0.047619	-1.562438	1.000000	2.500000
50%	0.142857	0.666667	0.142857	0.142857	0.095238	0.142857	-1.037803	2.380476	4.000000
75%	0.428571	0.857143	0.428571	0.380952	0.333333	0.333333	-0.266904	3.773340	5.500000
max	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000	0.996411	8.927174	11.181818

# 在这里先将df_feat_sample保存起来，为了研究过程中抽样结果的一致性。
# df_feat_sample.to_csv("df_feat_sample.csv" , encoding = 'gbk')

# read data from the pre-saved data
df_feat_sample = pd.read_csv("df_feat_sample.csv", encoding = 'gbk')
df_feat_sample.drop(columns=['Unnamed: 0'] , inplace = True)
df_feat_sample.head(5)

	CAR_NUM	F1	F2	O1	D1	O2	D2	FS	STD_FD	AVG_FD
0	浙AFZ391	0.047619	0.904762	0.142857	0.142857	0.571429	0.619048	-2.000000	2.093641	5.125000
1	浙B732UE	0.047619	0.428571	0.047619	0.047619	0.047619	0.047619	-1.453860	1.118034	2.333333
2	浙AN0G37	0.428571	0.952381	0.523810	0.523810	0.523810	0.523810	-0.677038	2.547726	4.909091
3	浙A3WA22	0.047619	0.571429	0.047619	0.047619	0.285714	0.333333	-1.199703	2.958040	3.666667
4	浙GN820X	0.476190	0.857143	0.476190	0.285714	0.285714	0.523810	-0.820671	5.780846	5.727273

Dimension reduction

In this part, i will do the dimension reduction by the factor analysis.

Choosing the Number of Factors

# Create factor analysis object and perform factor analysis
fa = FactorAnalyzer(n_factors = 9, rotation='varimax', method='principal', use_smc=True, \
                    is_corr_matrix=False, bounds=(0.005, 1), impute='median', rotation_kwargs=None)
fa.fit(df_feat_sample[['F1' , 'F2' , 'O1' , 'D1' , 'O2' , 'D2' , 'FS' , 'STD_FD' , 'AVG_FD']])
# Check Eigenvalues
factor_ev = fa.get_eigenvalues()
factor_ev

(array([4.73873204, 1.8188256 , 0.9094108 , 0.55263826, 0.3704023 ,
        0.24765782, 0.16004038, 0.12240567, 0.07988715]),
 array([4.73873204, 1.8188256 , 0.9094108 , 0.55263826, 0.3704023 ,
        0.24765782, 0.16004038, 0.12240567, 0.07988715]))

# Create scree plot using matplotlib
plt.rcParams['font.sans-serif'] = ['SimHei']  #显示中文字符
plt.rcParams['axes.unicode_minus']=False  # 用来正常显示负号
plt.rcParams['figure.dpi'] = 100 #图片像素
plt.scatter(range(1,df_feat_sample.shape[1]),factor_ev[1])
plt.plot(range(1,df_feat_sample.shape[1]),factor_ev[1])
plt.title('Scree Plot')
plt.xlabel('Components')
plt.ylabel('Eigenvalue')
plt.grid()
plt.show()

# Calculate the factor variance information, including variance, proportional variance and cumulative variance for each factor.
# 1. Sum of squared loadings (variance)
# 2. Proportional variance
# 3. Cumulative variance
factor_variance = fa.get_factor_variance()
factor_variance

(array([2.45404171, 2.11969221, 1.11573775, 1.03857742, 0.91038504,
        0.90374638, 0.25355476, 0.118127  , 0.08613773]),
 array([0.2726713 , 0.23552136, 0.12397086, 0.11539749, 0.10115389,
        0.10041626, 0.02817275, 0.01312522, 0.00957086]),
 array([0.2726713 , 0.50819266, 0.63216352, 0.74756101, 0.8487149 ,
        0.94913117, 0.97730392, 0.99042914, 1.        ]))

# see the factors' loadings
pd.DataFrame(fa.loadings_)

	0	1	2	3	4	5	6	7	8
0	0.642595	0.520488	0.039392	0.141357	0.058061	0.224747	0.490657	-0.000078	-0.001993
1	0.401459	0.359925	0.092731	0.033167	0.159006	0.818804	0.062139	0.001323	0.000408
2	0.892432	0.320055	-0.016298	0.099669	-0.010793	0.218011	0.041187	-0.018023	-0.203135
3	0.886677	0.339429	-0.043627	0.124019	-0.038802	0.185638	0.020372	0.021384	0.210873
4	0.383513	0.857136	-0.057268	0.107485	-0.034716	0.192692	0.048660	0.249954	0.014937
5	0.363498	0.866686	-0.028835	0.110198	-0.013491	0.212216	0.059132	-0.234223	-0.013214
6	0.130681	0.116006	0.004334	0.983743	-0.018573	0.025438	0.026511	0.000522	0.000880
7	-0.021097	-0.023752	0.459091	-0.025964	0.878569	0.124302	0.014618	-0.001459	-0.001407
8	-0.024013	-0.041749	0.942615	0.007729	0.326333	0.050413	0.007560	-0.001789	-0.001267

so, I choose 3 factors based on the principal of eigenvalues greater then 1.

Performing Factor Analysis

# Create factor analysis object and perform factor analysis using 5 factors
fa = FactorAnalyzer(n_factors = 3, rotation='varimax', method='principal', use_smc=True, \
                    is_corr_matrix=False, bounds=(0.005, 1), impute='median', rotation_kwargs=None)
fa.fit(df_feat_sample[['F1' , 'F2' , 'O1' , 'D1' , 'O2' , 'D2' , 'FS' , 'STD_FD' , 'AVG_FD']])
fa.loadings_

array([[ 0.9117434 ,  0.06458854,  0.15205703],
       [ 0.80155861,  0.28493596, -0.04960228],
       [ 0.89217186, -0.03025734,  0.10614231],
       [ 0.88641902, -0.07932703,  0.13627148],
       [ 0.87323388, -0.10114568,  0.10527671],
       [ 0.87187984, -0.05806726,  0.10580668],
       [ 0.18535326, -0.00653933,  0.97736579],
       [ 0.02614182,  0.93052244, -0.05183417],
       [-0.03774824,  0.9181832 ,  0.03922065]])

x = fa.loadings_
x = pd.DataFrame(x)
x

	0	1	2
0	0.911743	0.064589	0.152057
1	0.801559	0.284936	-0.049602
2	0.892172	-0.030257	0.106142
3	0.886419	-0.079327	0.136271
4	0.873234	-0.101146	0.105277
5	0.871880	-0.058067	0.105807
6	0.185353	-0.006539	0.977366
7	0.026142	0.930522	-0.051834
8	-0.037748	0.918183	0.039221

# Calculate the factor variance information, including variance, proportional variance and cumulative variance for each factor.
# 1. Sum of squared loadings (variance)
# 2. Proportional variance
# 3. Cumulative variance
factor_variance = fa.get_factor_variance()
factor_variance

(array([4.61465757, 1.81514586, 1.037165  ]),
 array([0.51273973, 0.20168287, 0.11524056]),
 array([0.51273973, 0.7144226 , 0.82966316]))

# see the eigenvalues
fa.get_eigenvalues()

(array([4.73873204, 1.8188256 , 0.9094108 , 0.55263826, 0.3704023 ,
        0.24765782, 0.16004038, 0.12240567, 0.07988715]),
 array([ 4.54198454e+00,  1.67073505e+00,  8.87829058e-01,  3.46499536e-01,
         1.60404388e-01,  6.65621855e-02,  1.67424546e-03, -8.99107766e-02,
        -1.18809797e-01]))

Get the new features' score

df_feat_new = fa.transform(df_feat_sample[['F1' , 'F2' , 'O1' , 'D1' , 'O2' , 'D2' , 'FS' , 'STD_FD' , 'AVG_FD']])
df_feat_new = pd.DataFrame(df_feat_new)
df_feat_new['CAR_NUM'] = df_feat_sample['CAR_NUM']
df_feat_new.rename(columns = {0 : 'f1' , 1 : 'f2' , 2 : 'f3'} , inplace = True)
df_feat_new.head(5)

	f1	f2	f3	CAR_NUM
0	0.633961	0.057235	-1.165552	浙AFZ391
1	-0.878013	-0.763536	-0.307468	浙B732UE
2	1.312621	0.156470	0.043351	浙AN0G37
3	-0.335299	-0.023774	-0.200037	浙A3WA22
4	0.847105	1.263833	-0.014173	浙GN820X

df_feat_new.describe()

	f1	f2	f3
count	4.623000e+03	4.623000e+03	4.623000e+03
mean	-5.792468e-17	-2.264634e-17	2.207238e-16
std	1.000108e+00	1.000108e+00	1.000108e+00
min	-1.408927e+00	-1.836644e+00	-3.331709e+00
25%	-8.057797e-01	-7.140758e-01	-5.158000e-01
50%	-3.135497e-01	-5.802492e-02	-4.925415e-02
75%	6.600085e-01	6.362371e-01	6.788130e-01
max	3.262468e+00	3.412270e+00	2.275098e+00

Understand the new factors

# plot the box plot, so that we can see the distribution of all the features.
fig,(ax1, ax2, ax3) = plt.subplots(1, 3, sharex = True,figsize=(15, 5))
# ax1 = plt.subplot(3, 3, 1,figsize=(10, 11))
ax1.violinplot(df_feat_new["f1"] )
ax1.set_xlabel('Factor 1')
# # ax2 = plt.subplot(3, 3, 2)
ax2.violinplot(df_feat_new["f2"] )
ax2.set_xlabel('Factor 2')
ax3.violinplot(df_feat_new["f3"] )
ax3.set_xlabel('Factor 3')
# plt.savefig('子样本车辆分布情况_new.png')
plt.show()

#将降维后的3个因子可视化展示出来
plt.rcParams['figure.dpi'] = 150 #分辨率
fig = plt.figure()
ax = Axes3D(fig)
ax.scatter(df_feat_new['f1'] , df_feat_new['f2'] , df_feat_new['f3'] , s = 1, c=None )
ax.set_xlabel("Factor 1") ; ax.set_ylabel("Factor 2") ; ax.set_zlabel("Factor 3") 
plt.show()

Commuting Pattern recognition

From then on, I start to get the commuting vehicles. The general process is clustering ths sample firstly, then train the decision tree model to get the commuting rule.

Get the optimal number of the clusters

Here i try to get the optimal cluster number by the gap statistic method.

# 利用本地的GAP_STATISTICS包里的代码得到gap(k)随着聚类数目k的改变而变化的图
from GAP_STATISTICS import gapStat
plt.rcParams['figure.dpi'] = 300 #分辨率
gaps , gapsDiff = gapStat(df_feat_new[['f1' , 'f2' , 'f3']].values)
print("gaps:" , gaps)
print("gaps_diff:",gapsDiff)

# plot the gap(k) and the Gap(k)-(Gap(k+1)-sk+1) Varying with the Number of Clusters
x_axis = range(1 , 10)
f, (a1,a2) = plt.subplots(2,1)
a1.plot(x_axis , gaps, 'g-o')
a1.set_ylabel(r'$Gap_n(k)$')
a2.bar(np.arange(1 , len(gapsDiff) + 1),gapsDiff)
a2.set_xlabel("Number of Clusters(k)")
a2.set_ylabel(r'$Gap_n(k)-(Gap_n(k+1)-s_{k+1})$')
f.show()

E:\Anaconda\lib\site-packages\scipy\cluster\vq.py:574: UserWarning: One of the clusters is empty. Re-run kmeans with a different initialization.
  warnings.warn("One of the clusters is empty. "

gaps: [0.44715245 0.43314783 0.45948223 0.44319516 0.41633712 0.40196826
 0.40662059 0.39496477 0.37610097]
gaps_diff: [ 0.02486359 -0.00571009  0.02299478  0.03104133  0.01768958 -0.00178513
  0.02403709  0.03202931]

E:\Anaconda\lib\site-packages\ipykernel_launcher.py:16: UserWarning: Matplotlib is currently using module://ipykernel.pylab.backend_inline, which is a non-GUI backend, so cannot show the figure.
  app.launch_new_instance()

DBSCAN

I have got the optimal clustering number that is 3. From then on, i will start to use several clustering method to identify the commuting vehicles. Firstly, i would like to try the DBSCAN method.

The initial parameters of DBSCAN determination

# 根据周红芳的论文中的方法来确定DBSCAN的2个初始参数.
# Firstly, I would like to get the eps
from DBSCANRelatedFunction import GetEps
# importlib.reload(DBSCANRelatedFunction)
eps_array = GetEps(df = df_feat_new[['f1' , 'f2' , 'f3']])
print("The fourth eps in the eps array is : " , eps_array[4])

The fourth eps in the eps array is :  0.2143134550456535

# secondly, I would like to get the minPts
from DBSCANRelatedFunction import GetMinPts
# importlib.reload(DBSCANRelatedFunction)
minPts = GetMinPts(df = df_feat_new[['f1' , 'f2' , 'f3']] , eps = eps_array[4])
print("The best minPts is : " , minPts)

The best minPts is :  7.767683322517845

DBSCAN Clustering

from DBSCANRelatedFunction import DBSCAN
# importlib.reload(DBSCANRelatedFunction)
labels = DBSCAN(df = df_feat_new[['f1' , 'f2' , 'f3']] , eps_set = 0.2143 , MinPts = 8 )

每个样本的簇标号:
[ 1  0 -1 ... -1 -1  0]
噪声比: 47.67%
分簇的数目: 34
轮廓系数: -0.247

# Analyse the result of the DBSCAN clustering
df_feat_new['DBSCAN'] = labels
# 其中聚类结果中最大的簇为第0个簇，有2178辆车，得到该簇的统计信息
df_feat_new = pd.merge(df_feat_new , df_feat_sample , on = ['CAR_NUM'] , how = 'inner')
df_feat_new.head(5)

	f1	f2	f3	CAR_NUM	DBSCAN	F1	F2	O1	D1	O2	D2	FS	STD_FD	AVG_FD
0	0.633961	0.057235	-1.165552	浙AFZ391	1	0.047619	0.904762	0.142857	0.142857	0.571429	0.619048	-2.000000	2.093641	5.125000
1	-0.878013	-0.763536	-0.307468	浙B732UE	0	0.047619	0.428571	0.047619	0.047619	0.047619	0.047619	-1.453860	1.118034	2.333333
2	1.312621	0.156470	0.043351	浙AN0G37	-1	0.428571	0.952381	0.523810	0.523810	0.523810	0.523810	-0.677038	2.547726	4.909091
3	-0.335299	-0.023774	-0.200037	浙A3WA22	0	0.047619	0.571429	0.047619	0.047619	0.285714	0.333333	-1.199703	2.958040	3.666667
4	0.847105	1.263833	-0.014173	浙GN820X	-1	0.476190	0.857143	0.476190	0.285714	0.285714	0.523810	-0.820671	5.780846	5.727273

# see the biggest cluster's statistical infomation
df_feat_new[df_feat_new['DBSCAN'] == 0].describe()

	f1	f2	f3	DBSCAN	F1	F2	O1	D1	O2	D2	FS	STD_FD	AVG_FD
count	2056.000000	2056.000000	2056.000000	2056.0	2056.000000	2056.000000	2056.000000	2056.000000	2056.000000	2056.000000	2056.000000	2056.000000	2056.000000
mean	-0.509467	-0.184230	-0.080041	0.0	0.142510	0.488327	0.142880	0.128822	0.124189	0.136395	-1.183317	2.218592	3.710175
std	0.605167	0.815186	0.465648	0.0	0.142020	0.270258	0.150280	0.137450	0.121399	0.128306	0.458078	1.599549	1.989572
min	-1.364080	-1.779969	-1.185855	0.0	0.047619	0.047619	0.047619	0.047619	0.047619	0.047619	-2.464318	0.000000	0.000000
25%	-0.952084	-0.769789	-0.357086	0.0	0.047619	0.238095	0.047619	0.047619	0.047619	0.047619	-1.440093	0.957427	2.400000
50%	-0.681525	-0.161238	-0.113528	0.0	0.095238	0.476190	0.095238	0.047619	0.095238	0.095238	-1.141871	2.163457	3.600000
75%	-0.249902	0.429047	0.143619	0.0	0.190476	0.714286	0.142857	0.142857	0.142857	0.142857	-1.000000	3.354917	5.000000
max	1.781952	1.700654	2.139364	0.0	0.904762	1.000000	0.952381	0.904762	0.857143	0.809524	0.990016	7.204165	11.000000

ISODATA Clustering

Get the initial parameters of ISODATA by the kmeans

import k_means_cluster
from max_min_cluster import get_distance
result = k_means_cluster.k_means_cluster(df_feat_new[['f1' , 'f2' , 'f3']].values , 3)
result

[[array([ 0.63396088,  0.05723516, -1.16555231]),
  array([ 0.84710534,  1.26383289, -0.01417255]),
  array([-0.37747135,  0.11295772, -0.15800405]),
  array([0.72452982, 1.53138035, 1.70853138]),
  array([-0.45757533,  0.92513296,  0.82505124]),
  array([-0.04388197,  0.42247264, -0.0647826 ]),
  array([-0.89865666,  2.5771406 ,  0.18062256]),
  array([-0.6429409 ,  0.62700037, -0.2861031 ]),
  array([ 0.38066193,  0.12140998, -1.09114632]),
  array([ 0.38993321,  0.47727149, -0.45567808]),
  array([-0.55689067,  0.58459735,  0.01096489]),
  array([-0.65807526,  0.67965687, -0.15508679]),
  array([0.21332963, 0.83853658, 0.12939905]),
  array([-0.25361311,  1.31215145, -0.36571314]),
  array([-0.80636449,  0.64907786,  0.04485982]),
  array([-0.59744551,  0.81517692,  0.8404408 ]),
  array([-0.58711568,  0.92494758,  0.35568915]),
  array([ 0.51983573,  0.50231169, -0.338939  ]),
  array([-0.27275897,  0.14749072, -0.2574021 ]),
  array([-0.06945247,  1.23296242, -0.1122991 ]),
  array([0.32889894, 1.80571778, 1.50864109]),
  array([ 0.46809367,  0.20947661, -0.35479834]),
  array([-0.77712875,  1.11165073, -2.42404474]),
  array([-0.75689946,  1.56571406, -1.07141465]),
  array([ 0.62308757,  1.31377211, -0.12613174]),
  array([-0.82423948, -0.24299913, -2.59445645]),
  array([ 1.05118604,  0.32431144, -2.37184458]),
  array([ 0.50696437,  0.08330056, -1.00025714]),
  array([ 1.0405091 , -0.32026518, -3.12167278]),
  array([ 0.34122186, -0.04929885, -0.33277382]),
  array([-0.87868178,  1.31493092, -0.03964988]),
  array([-0.69067172,  0.06865871, -2.60327721]),
  array([ 0.14485893,  0.03443066, -1.6379793 ]),
  array([0.66223742, 1.11377332, 0.39089227]),
  array([-0.44572616,  0.08066976, -0.38259466]),
  array([ 0.12286555,  1.76285734, -0.08442848]),
  array([-0.81088048,  0.71022538,  0.36307957]),
  array([ 0.24165062,  0.07494784, -0.33122391]),
  array([-0.64742398,  2.83183036, -0.42720924]),
  array([1.52961534, 2.07734384, 0.31706005]),
  array([0.10369686, 0.55179571, 1.18887442]),
  array([-0.78452976,  2.81345768, -0.68091075]),
  array([-0.56931382,  1.46407498,  0.05540931]),
  array([-1.34210914,  1.20094823,  1.93798901]),
  array([ 0.81737966,  1.72691732, -0.43242197]),
  array([ 0.44201242,  2.73867927, -0.11523875]),
  array([ 0.19509555,  1.28264383, -0.86979563]),
  array([-1.02029603,  0.92039638,  0.23357188]),
  array([-0.24986603,  0.66259646,  0.0405629 ]),
  array([-0.58102653, -0.11213735, -2.59787536]),
  array([-0.07107469,  0.57585395, -0.85384191]),
  array([-0.61882787,  0.32095022, -0.58974334]),
  array([ 0.70338373,  0.89621685, -0.12201239]),
  array([0.3733395 , 1.84762808, 0.2506463 ]),
  array([-0.93123587,  0.59842728,  0.55733315]),
  array([-0.73976588,  0.6524291 , -1.26589343]),
  array([-0.38505305,  1.73947163,  0.09744304]),
  array([-0.413967  ,  0.22927787, -0.01297715]),
  array([-0.74772041,  0.11507082, -0.55141212]),
  array([-0.9497187 ,  1.13364209, -0.02084769]),
  array([-0.95484379,  2.73325991, -0.89407628]),
  array([-0.28146225,  0.27882517, -0.072536  ]),
  array([-0.09742558, -0.04621289, -1.23999913]),
  array([0.18295592, 0.34482748, 0.63525923]),
  array([-0.3758128 ,  0.16163198, -0.12619594]),
  array([1.06520168, 2.52785835, 1.83378431]),
  array([-1.19726016,  2.04868584,  0.51836966]),
  array([-1.03779242,  0.54847839,  0.02754745]),
  array([ 0.19863451,  0.29364223, -0.30349819]),
  array([-0.03322939, -0.29334731, -1.17429067]),
  array([-0.28703413, -0.21131176, -1.84265684]),
  array([-0.46739587, -0.09826034, -1.98010891]),
  array([-0.26542217,  0.0598723 , -1.85046963]),
  array([-0.33790458,  1.11330755,  0.63031933]),
  array([-0.6747734 ,  0.99146585, -2.24733969]),
  array([-0.58029805,  0.63588705,  0.24118933]),
  array([ 1.35985661,  1.72008923, -0.22000916]),
  array([-0.34863475, -0.27460455, -3.010665  ]),
  array([-0.03176683,  1.72516451, -0.10686867]),
  array([-0.67334067,  1.2793974 , -0.92241323]),
  array([ 0.16040129,  1.22678822, -0.44053944]),
  array([-0.15919496,  0.77607959,  0.29348516]),
  array([-0.12789552,  1.89890761, -0.30595454]),
  array([-1.03774008,  1.11386585,  1.27145205]),
  array([-0.95460146,  0.74497747, -0.51663527]),
  array([-0.12552242,  0.62841487, -0.07665992]),
  array([-0.9325524 ,  2.86104014, -0.50910166]),
  array([-0.78472887,  0.41173623, -0.0823075 ]),
  array([ 0.33945766,  0.17048092, -0.89962393]),
  array([-0.45193021,  1.71023238, -1.84167693]),
  array([-0.2933035 ,  0.5802775 , -0.00204659]),
  array([-0.37223045,  2.80184173,  0.05055469]),
  array([ 0.96085011,  2.27826258, -0.20529275]),
  array([-0.43046516,  2.5123601 , -0.72293717]),
  array([-0.66413377,  0.9902174 , -0.19215231]),
  array([-0.44181662,  0.88753613,  0.12915311]),
  array([ 0.16766571,  0.271743  , -0.11885742]),
  array([ 0.1013477 ,  0.46123549, -0.76475528]),
  array([-1.15924799,  1.35889341,  1.20297361]),
  array([-0.6757842 ,  0.58079053,  0.98375573]),
  array([-0.89254494,  0.64557681, -1.06762817]),
  array([-0.59835994,  2.36233455,  0.13355866]),
  array([-0.73011726,  1.75803925, -0.34986387]),
  array([-0.3268377 , -0.11202001, -1.27352324]),
  array([0.77723714, 2.32820841, 0.37069899]),
  array([-0.5967295 ,  2.90367266,  0.15011029]),
  array([-0.97918774,  1.11815314, -0.44837158]),
  array([-0.12526515,  1.74633404, -0.05036957]),
  array([-0.5246856 ,  1.05669729,  0.11952708]),
  array([-0.1673414 ,  1.95384025,  0.57865935]),
  array([0.76434982, 1.4445013 , 1.39648763]),
  array([-1.04614862,  0.58009518, -0.62667787]),
  array([ 0.10562381,  0.47034812, -0.37651914]),
  array([-0.96015   ,  2.00646522,  0.35290979]),
  array([-0.8210447 ,  1.73554655,  0.79294939]),
  array([ 0.67631806,  1.02247504, -0.13671026]),
  array([-0.86187869,  1.21025076, -1.0660815 ]),
  array([-0.95180835,  0.99474129, -0.28156134]),
  array([-0.71903865,  0.05795438, -0.51235814]),
  array([-0.2345421 ,  0.28928248,  0.18830146]),
  array([-0.12410729,  0.48622325,  0.5542355 ]),
  array([ 0.6702066 ,  0.49005358, -0.18891407]),
  array([-0.67399855,  0.12287347, -0.42428236]),
  array([0.73731152, 1.55661602, 0.30817377]),
  array([-0.58179284,  0.72863789, -1.41275997]),
  array([-0.43607028,  1.39728938,  0.42109146]),
  array([-0.90779878,  0.81120662,  0.56076097]),
  array([-0.96725379,  1.25614412,  0.0550662 ]),
  array([-0.64484027,  1.18932399, -0.22022999]),
  array([-0.74297293,  0.17570337, -0.35717508]),
  array([-0.90857427,  0.29175566, -0.02449692]),
  array([ 0.48594276,  2.17900353, -0.59704138]),
  array([-0.41517825,  0.47553079,  0.22266643]),
  array([-1.2557654 ,  1.04114252,  1.49878711]),
  array([ 0.51107457,  0.72563082, -2.79281049]),
  array([-0.57702443,  1.43302308, -0.05203693]),
  array([-0.08202384,  1.02972901,  0.16662698]),
  array([-0.37306586,  0.83580602, -0.27192468]),
  array([-0.59952805,  0.35656944, -0.10673124]),
  array([ 1.49581273,  1.01663343, -1.40286258]),
  array([-0.31280746, -0.00838799, -0.8243342 ]),
  array([0.81530998, 1.56160957, 1.243693  ]),
  array([0.1860417 , 2.22266624, 0.03650052]),
  array([ 0.01828335, -0.10737264, -1.41147161]),
  array([-0.43965254, -0.26438528, -2.52952659]),
  array([-0.21481214,  0.7440546 , -0.14496936]),
  array([-1.12635561,  1.44138366,  1.71254868]),
  array([-0.21201527,  0.46455224, -0.87178764]),
  array([-0.65811767,  0.84923488, -1.92183613]),
  array([-0.50876207,  0.67962787, -1.80721969]),
  array([-0.88779753,  0.45280157,  0.27371796]),
  array([-0.47302572,  0.8204095 ,  0.1396317 ]),
  array([ 0.22635734,  0.81430295, -0.3467514 ]),
  array([-0.66366687,  1.13103522, -0.0210606 ]),
  array([ 0.17755585,  2.46416114, -0.07224631]),
  array([-0.20887264,  1.04440012, -0.45739742]),
  array([-0.65328351,  0.03448757, -0.70972124]),
  array([-0.78484262,  0.97058629, -0.81355097]),
  array([ 0.71805909, -0.30104246, -2.8140686 ]),
  array([-0.80715373,  1.25029499, -0.38329972]),
  array([-1.29543169,  1.98222951,  1.90791952]),
  array([-0.22566121,  1.74616827, -0.61970095]),
  array([-1.0340615 ,  0.61841898, -0.07244919]),
  array([ 0.40497856,  1.61200947, -1.26060057]),
  array([-0.65709378,  0.68754541, -0.59433748]),
  array([-0.27772788,  0.90543809, -0.09456491]),
  array([ 0.09601871,  0.09328567, -0.07034502]),
  array([-1.07986962,  0.66519362,  0.06215609]),
  array([-0.79889459,  1.05921032, -0.01644525]),
  array([0.54916117, 0.85081858, 0.66440462]),
  array([-0.60122571,  1.34410247,  1.59127141]),
  array([-0.61299899,  0.15279458, -0.38320098]),
  array([-1.08765614,  0.39519503, -0.36833156]),
  array([-0.24160139,  0.41322282, -2.85981504]),
  array([-0.36029717,  1.22505561, -0.3467933 ]),
  array([-0.61922372,  0.37190968, -0.14385379]),
  array([-0.02463308,  1.49844574, -2.07835152]),
  array([-1.04823867,  0.83872746,  0.51938504]),
  array([-0.45758796,  0.68633967, -1.35602503]),
  array([ 0.88536162, -0.40499963, -3.33170906]),
  array([-0.32676847,  1.42007859,  0.57471379]),
  array([-0.52774687,  0.84912453, -0.36608384]),
  array([ 0.42482332,  1.30997732, -1.03117786]),
  array([-1.08849025,  1.61971344,  0.87849204]),
  array([-0.5743301 ,  1.98984901,  0.14602829]),
  array([-0.80832293,  0.32297295, -1.15257785]),
  array([-0.2451756 ,  0.60263656, -1.99754106]),
  array([-0.96598624,  0.56179977,  0.7644227 ]),
  array([-0.58317898,  1.120928  ,  1.19890794]),
  array([-0.21488039,  0.17422363, -1.36110226]),
  array([-0.9561683 ,  2.29080706, -0.16554629]),
  array([-0.44020596,  0.46139652,  0.50100962]),
  array([-0.6044895 ,  0.97326035,  0.21062466]),
  array([0.81831683, 1.73223296, 0.11054022]),
  array([-0.46983142,  0.56484756, -0.0193081 ]),
  array([-0.49531343,  0.17251312, -0.29694691]),
  array([-0.84687935,  0.73091661, -1.99536956]),
  array([-0.54993272,  2.15935903,  0.02484269]),
  array([-0.6361508 ,  0.14242256, -0.25049263]),
  array([-0.61809296,  0.47488996, -0.49376027]),
  array([-0.1507817 ,  2.3039871 , -0.49175778]),
  array([-0.11517097,  0.65682234, -0.20023406]),
  array([-0.74103478,  2.48202449, -0.2929216 ]),
  array([ 0.14866324,  0.78568107, -1.50281202]),
  array([-0.24487446,  0.48904035, -0.32657991]),
  array([-0.55285519,  1.85579092,  1.7150592 ]),
  array([-0.91450722,  1.99538942,  0.45105612]),
  array([0.66905665, 1.4847622 , 1.1301676 ]),
  array([ 0.20654793,  0.3348906 , -0.28853553]),
  array([-0.65270715,  1.11772936,  1.86458273]),
  array([-0.58316141,  0.66072113,  0.57562198]),
  array([-0.85245135,  1.63519255, -0.29835794]),
  array([-0.84973007,  0.06684647, -0.84434338]),
  array([-1.10133667,  0.76564761,  0.45870675]),
  array([-1.12975086,  0.55408765,  0.00658619]),
  array([-1.11337827,  1.13717377,  0.95110075]),
  array([-0.97189739,  0.62424915,  0.31371956]),
  array([-0.38749243, -0.0668971 , -1.86185008]),
  array([-0.60196377,  2.91909087, -0.4324728 ]),
  array([-0.48525699,  1.335705  , -1.57801625]),
  array([-0.560698  ,  0.77570443,  0.92788779]),
  array([1.63808318, 2.82007455, 0.90894712]),
  array([-0.7274656 ,  1.22199854, -0.3826805 ]),
  array([ 0.1075178 ,  0.65983633, -0.59151906]),
  array([-0.13830164,  0.88618197, -2.13265578]),
  array([-0.69707189,  1.9009344 ,  1.37653863]),
  array([-0.90960165,  1.71042878, -1.86103588]),
  array([-0.99772474,  0.5855096 , -0.66954563]),
  array([-0.34491303,  0.97527528, -0.11703818]),
  array([-0.73795346,  0.67220561, -0.44070743]),
  array([ 0.06809413,  0.23649407, -1.00892396]),
  array([-0.22416668,  0.21210937, -0.55667044]),
  array([ 0.45785503,  0.61502471, -0.04926734]),
  array([-0.54540177,  0.46285326, -0.39234102]),
  array([ 0.55633098,  0.24408441, -0.6428233 ]),
  array([-0.76552952,  1.09371621, -0.24398794]),
  array([-0.90328474,  0.36802603, -0.09425458]),
  array([-1.11870142,  1.34479158,  0.30841308]),
  array([ 0.55129887,  0.71517098, -0.23814632]),
  array([-0.10594466,  0.62454056, -0.32699178]),
  array([-0.18598775,  0.30556948, -1.01488653]),
  array([-0.46966303,  1.25245133, -0.22748556]),
  array([-0.49158486,  1.16483903, -0.30640474]),
  array([ 1.05086281,  0.61699455, -1.60644404]),
  array([ 0.55127425,  1.63572552, -0.15796227]),
  array([-0.10983068,  0.28519594, -0.05717501]),
  array([-0.15136883,  0.46499332, -0.98095595]),
  array([-0.39301906,  0.09021091, -0.42406107]),
  array([-0.38127174,  1.54206276,  1.47115485]),
  array([-0.07491242,  1.1632568 , -0.11549595]),
  array([-0.79646594,  0.99120835, -0.75868233]),
  array([-0.71738042,  0.20520676, -0.39718778]),
  array([-0.74955715,  1.56158767,  0.0623632 ]),
  array([-0.96745907,  1.90852993, -1.66174   ]),
  array([-0.80589733,  0.29030729, -0.85307072]),
  array([0.90566715, 2.16369904, 0.01196258]),
  array([0.17868786, 0.98278822, 1.2140487 ]),
  array([-0.92375707,  2.3108067 ,  0.32402769]),
  array([-0.81951115,  1.96668746,  1.32280551]),
  array([-0.68128456,  0.60070941, -0.7540157 ]),
  array([-0.25090396,  0.08932615, -0.4182632 ]),
  array([-0.69954253,  1.06893979,  0.21995365]),
  array([-0.47900926,  1.03064686,  0.11837322]),
  array([-0.32436855,  0.90949632, -0.07798186]),
  array([-0.60376245,  1.33622332, -0.33670614]),
  array([ 0.74576421,  2.4241983 , -0.53928589]),
  array([-0.62150441,  0.87773209,  1.54898032]),
  array([-1.0675884 ,  1.79423688,  0.3850528 ]),
  array([ 0.75158882,  0.40654504, -0.99383086]),
  array([ 0.15123403,  0.61711534, -0.0609659 ]),
  array([-0.29932625,  0.1588194 , -2.99481466]),
  array([-0.03977771,  0.05414323, -0.26991056]),
  array([-0.09067072,  2.84047485,  0.09366543]),
  array([-0.62335221,  1.4952485 , -0.20690088]),
  array([-0.79605896,  0.57814749,  0.22234547]),
  array([-1.0225788 ,  1.14961069, -0.76174315]),
  array([-1.13358923,  2.24908618,  0.06597929]),
  array([-0.35720132,  1.30248977, -0.26207919]),
  array([ 0.68156994,  0.77176732, -0.84055173]),
  array([-1.01792249,  0.55301998,  0.58289694]),
  array([-0.62544746,  2.88596406, -2.30855949]),
  array([-0.62909147,  0.68549234,  0.91535227]),
  array([-0.3061218 ,  0.76707958, -2.24529428]),
  array([0.01338156, 3.11153968, 1.09824192]),
  array([-0.88773138,  0.58875386, -1.38135535]),
  array([-1.1246076 ,  0.21303868, -0.3797401 ]),
  array([-0.90320932,  0.82561903,  0.26043598]),
  array([-0.94358365,  2.44670301, -1.69348505]),
  array([-1.02990233,  1.19350165,  1.16858992]),
  array([-0.58964176,  0.65134692,  0.12670697]),
  array([-0.19467513,  1.49824648,  0.12667877]),
  array([-0.05923445,  1.50081516, -0.31212515]),
  array([-0.69864634,  1.40898067,  0.16318536]),
  array([ 0.10320739,  0.63799676, -0.86936324]),
  array([-0.90448559,  0.44418794,  0.45225954]),
  array([-0.41312127,  0.28812158,  0.04915408]),
  array([ 0.15916507,  0.32935222, -0.26598683]),
  array([-0.46212287,  2.07435143, -0.4651529 ]),
  array([-0.74037636,  0.29438821, -0.53369351]),
  array([-0.50840491,  0.74311699,  0.23262587]),
  array([ 0.98514494,  1.8663363 , -0.11760417]),
  array([-0.22584676,  0.20187856, -0.17546346]),
  array([0.34344369, 2.22770891, 0.99049066]),
  array([-0.54082336,  2.21120839, -0.51413675]),
  array([-0.93503903,  0.91543411,  0.19273932]),
  array([0.28343295, 0.89838314, 0.83958933]),
  array([-0.60764427,  2.79928835, -0.04170019]),
  array([0.35019125, 0.17536782, 0.07524441]),
  array([-0.62992401,  0.94369715,  0.150232  ]),
  array([0.755665  , 1.04333233, 0.21453588]),
  array([ 0.09927544,  0.08531641, -2.19717276]),
  array([-0.19214126,  1.68755483,  1.88729742]),
  array([-0.92778024,  0.80906611, -0.23822084]),
  array([0.31650473, 0.64403038, 0.68675457]),
  array([ 0.54818187,  0.47021378, -1.08934417]),
  array([ 0.21371759,  0.94200564, -1.16824871]),
  array([0.37373432, 1.37435576, 1.61039202]),
  array([0.40606425, 1.06350207, 1.06826597]),
  array([ 0.09753314, -0.02602396, -0.21217382]),
  array([-0.62828676,  0.3237611 , -0.23710486]),
  array([-0.94150133,  1.3265883 ,  0.31118033]),
  array([ 0.85663135,  0.7388404 , -0.12719439]),
  array([ 0.0903744 ,  1.81349079, -1.65468339]),
  array([-0.09938821,  0.9231043 , -0.43272634]),
  array([-0.08023443, -0.25596644, -2.11838943]),
  array([-0.68384352,  1.29601257,  1.4723231 ]),
  array([-0.48218253,  1.15887741,  0.94092751]),
  array([-0.46826244,  0.55796454,  0.22814151]),
  array([-0.60864441,  0.45901349, -0.49249135]),
  array([ 0.85852352,  0.20723613, -1.1264894 ]),
  array([ 0.19842506,  0.16657795, -2.2167741 ]),
  array([0.26095868, 0.90804067, 1.35490086]),
  array([-0.10130614,  2.13133865, -0.22946956]),
  array([-0.541265  , -0.12167153, -2.87434631]),
  array([ 1.07558395,  0.68725114, -1.06959481]),
  array([-0.18835746, -0.23154381, -1.6281327 ]),
  array([ 0.64700272,  0.40864271, -0.44815696]),
  array([-0.42143976, -0.76540229, -2.97596587]),
  array([-0.48952563,  0.71094141, -2.63552306]),
  array([-0.88766074,  1.77389016, -0.36449231]),
  array([ 0.3425021 ,  1.69868792, -0.16225802]),
  array([-0.65932574,  1.36292441, -0.57320587]),
  array([0.26798164, 1.5744408 , 1.58929784]),
  array([-0.76693538,  1.14976622,  0.16068633]),
  array([-0.64439787,  0.70525375, -1.12493014]),
  array([-0.37522823,  2.35279458, -1.77093086]),
  array([ 2.31428517,  1.24669425, -3.0606129 ]),
  array([ 0.24636208,  0.71324928, -0.36605654]),
  array([-0.22254843,  0.56487936,  1.21613034]),
  array([-0.42123078,  0.92924636, -0.92714724]),
  array([-0.99370309,  2.25606461,  0.19008778]),
  array([-0.76358516,  1.05869807, -0.72970552]),
  array([-0.4364458 , -0.22851287, -1.13284589]),
  array([-0.33247902,  0.88916533, -0.25629407]),
  array([-0.8850387 ,  2.29622105,  0.20459231]),
  array([-0.20069967,  2.67081999, -0.30143202]),
  array([-0.19175249,  0.76532673, -0.53512296]),
  array([ 0.48675424,  0.18148776, -2.25118517]),
  array([-0.44486362,  1.68763303,  1.48043271]),
  array([0.19828836, 0.54196075, 0.20048856]),
  array([-0.60896515,  1.93664772, -0.10425992]),
  array([-0.85048588,  1.97941885,  0.94252877]),
  array([-0.60020682,  0.23007911, -0.8962006 ]),
  array([ 0.45044921,  1.8141439 , -0.1692828 ]),
  array([-0.40985801,  0.26573533, -0.89249301]),
  array([-1.0700224 ,  0.85225244,  0.16392486]),
  array([-0.67680757,  0.33490359, -0.37547922]),
  array([-0.48069605,  1.25848947, -0.69898997]),
  array([-0.12571334,  0.03910775, -0.43921933]),
  array([-0.03042795,  2.07273923, -0.05548975]),
  array([ 0.25399882,  2.3398021 , -0.39250619]),
  array([-0.47460157,  0.49988258, -0.48119511]),
  array([-9.88507015e-01,  1.09583316e+00, -7.51275557e-04]),
  array([-0.1992482 ,  0.95063741,  0.41415171]),
  array([ 0.14286336,  0.33217192, -0.86309555]),
  array([-0.47280295,  0.58086872,  1.05146157]),
  array([-0.07346728,  0.39001175, -0.02659537]),
  array([-1.115774  ,  2.63872784,  1.45133369]),
  array([-0.43010634,  0.55219012, -0.66990575]),
  array([-0.02700899,  0.915166  , -0.16961925]),
  array([-0.59247112,  1.16612505,  0.29960604]),
  array([0.14649248, 2.46270262, 1.33320867]),
  array([ 0.39801653,  0.19596617, -0.06928446]),
  array([-0.33026086, -0.13589247, -0.93090109]),
  array([-0.58041444,  0.79600107, -1.47289777]),
  array([-0.21410277,  0.05253095, -2.41732687]),
  array([-0.34122486,  2.42779349, -0.0629106 ]),
  array([-0.72387805,  1.44635215,  0.7747418 ]),
  array([-0.37163834,  1.18519145, -0.72682311]),
  array([-0.26142063, -0.21310924, -2.94613168]),
  array([-0.01847515,  0.96081223, -0.63889258]),
  array([-0.32675527,  0.32021581, -0.92128709]),
  array([-0.44137826, -0.00170383, -0.50682008]),
  array([-0.22734627,  0.33329263, -0.06585022]),
  array([-0.01243567,  1.5529546 ,  1.85677672]),
  array([-0.64544256,  0.06264791, -2.45006366]),
  array([-0.71540095,  1.31211262,  0.37580406]),
  array([ 0.03568226,  2.23222182, -0.16141247]),
  array([ 0.78110754, -0.20746657, -2.70592556]),
  array([ 0.28257502,  1.04273483, -0.81399115]),
  array([0.79934958, 1.89808414, 0.75033473]),
  array([-0.35907544,  0.92619068,  1.09800051]),
  array([-0.33818337, -0.07708191, -1.29776971]),
  array([-0.55422158,  1.17250087, -0.69682277]),
  array([-0.39437773,  0.90772853, -0.50546625]),
  array([ 0.54860495,  1.5948529 , -0.52140517]),
  array([-0.97660004,  1.97483483, -0.72944892]),
  array([-0.31381867,  0.57690899, -2.44020674]),
  array([-0.59801948,  0.73911374, -0.0212007 ]),
  array([1.420673  , 2.27931529, 0.73128378]),
  array([ 1.45395111,  1.62802636, -0.48838367]),
  array([-0.56955226,  0.83537922, -1.1021117 ]),
  array([-0.67840628,  1.05396429,  0.4732104 ]),
  array([-1.01353124,  1.23590542,  0.66617031]),
  array([-0.43832816,  1.92698878, -1.37045064]),
  array([-0.74018667,  2.40911956, -0.19944616]),
  array([-0.65252357,  1.58315362, -0.20731899]),
  array([-0.37153759, -0.29369232, -1.83796871]),
  array([0.2885116 , 1.22658613, 1.15276868]),
  array([ 0.46499568,  1.17436738, -0.30882423]),
  array([ 0.91696458,  2.49605024, -0.3056308 ]),
  array([-0.75840404,  1.30060373, -1.58203609]),
  array([-0.65901953,  0.7033862 ,  0.59462594]),
  array([ 0.73989465,  1.08919177, -0.02726913]),
  array([-0.83032803,  0.40567568, -0.01494836]),
  array([-0.79729725,  0.66401981,  0.28885127]),
  array([ 0.64389386, -0.48462856, -2.86706147]),
  array([-0.81283845,  0.95859581,  1.46379052]),
  array([-0.33380617,  0.79455289,  1.61878668]),
  array([0.02042471, 0.3911767 , 0.92603307]),
  array([-0.59735742,  0.57687959,  0.93241459]),
  array([ 0.25568025,  1.446396  , -0.06606639]),
  array([-1.00046333,  0.72282327, -0.74620342]),
  array([-0.9762962 ,  1.23793634,  1.14909242]),
  array([-0.89329027,  1.0505895 , -0.01084639]),
  array([-0.67406328,  0.43371512, -0.58558362]),
  array([-0.33651859, -0.08817914, -0.95870576]),
  array([-0.59793423,  1.13989353, -0.65305039]),
  array([-0.7601085 ,  2.70862244,  0.77642306]),
  array([-0.64453873,  2.87115462,  1.28544925]),
  array([ 0.33451629,  1.58870716, -2.01036589]),
  array([-0.51864426,  2.80002189, -0.21803457]),
  array([-0.43042568,  1.95713806, -1.28589345]),
  array([-0.39396197,  3.38901267,  0.01600562]),
  array([ 0.29389058,  0.01327699, -0.33915269]),
  array([ 0.00522895,  0.89142102, -2.3582198 ]),
  array([-0.83886477,  1.16638223,  0.27301431]),
  array([-0.90927412,  1.44669931,  0.32847664]),
  array([-0.24393418,  0.75680417,  1.21146217]),
  array([ 0.76394171,  1.24820563, -0.14724742]),
  array([ 0.01388208,  0.15381363, -1.48448534]),
  array([0.62651383, 1.33211853, 0.1612679 ]),
  array([-1.03160879,  1.15199888, -0.67762332]),
  array([-0.47461773, -0.30784411, -2.89136178]),
  array([-0.43664963,  0.61467121, -1.87870285]),
  array([-0.53895674,  1.01716475,  0.02526325]),
  array([ 0.01656775,  0.69476066, -1.18011292]),
  array([0.4388848 , 1.30231629, 1.46603772]),
  array([ 0.32440033,  2.11901463, -0.73941046]),
  array([ 0.66929952,  1.22299685, -0.42990977]),
  array([0.63241068, 2.47568128, 0.94295391]),
  array([-0.3660282 ,  0.87967369, -0.92296138]),
  array([-0.62999912,  1.04413378, -0.28206665]),
  array([ 0.09520446,  1.05018304, -2.45235765]),
  array([-0.24230636,  0.48417871, -0.83537221]),
  array([-0.07348146,  1.33607639, -0.18549531]),
  array([ 0.07009096,  0.58584394, -0.22535748]),
  array([-0.25697612,  0.18730037, -2.74432189]),
  array([-0.76371827,  0.78915046,  0.1810409 ]),
  array([-0.80018759,  1.40685951,  1.62790627]),
  array([-0.6259987 ,  0.38711084,  0.12089484]),
  array([-0.50285381,  0.51686127,  0.10671403]),
  array([-0.7790312 ,  0.81514639, -0.70569663]),
  array([-0.34010121,  1.85547685,  0.50237818]),
  array([ 0.02951361, -0.02594592, -0.76452438]),
  array([-0.83758099,  0.45889641, -0.15825771]),
  array([-1.10257067,  0.43201703, -0.18000113]),
  array([ 0.07841419,  1.47502756, -0.51616103]),
  array([ 1.09110334,  0.20786783, -2.97638927]),
  array([-0.60719599,  0.96816397, -0.2478897 ]),
  array([-0.25944429,  0.73143705,  0.96196898]),
  array([-0.34353142,  1.71117155, -0.29399282]),
  array([-0.67736953,  2.68813307, -0.56459416]),
  array([-0.64212438,  0.73413584,  0.37949882]),
  array([ 0.38500405,  0.65204354, -0.25841052]),
  array([-0.91317255,  0.79010413,  0.78193513]),
  array([ 0.66674871,  1.46010841, -1.52530178]),
  array([-0.63817504,  1.06296766,  0.07479702]),
  array([-0.5498841 ,  1.91469171,  0.02912414]),
  array([ 0.75013083,  0.80731419, -0.07865385]),
  array([-0.63970341,  0.74753783, -0.15875868]),
  array([-0.80852298,  1.63483515,  0.20761521]),
  array([0.42556086, 1.42190852, 0.12842131]),
  array([-0.17817015,  1.80910097, -0.15496936]),
  array([ 0.31509969,  0.82409569, -0.24807836]),
  array([ 0.56208426,  0.74408023, -0.20038844]),
  array([-0.70402728,  1.35065675, -0.23224407]),
  array([-0.20846814,  0.49183463, -0.08560965]),
  array([-0.75176709,  0.75902228,  0.35139692]),
  array([ 0.38979575,  0.84550727, -0.15933629]),
  array([-0.86785758,  1.91681154, -0.05612459]),
  array([-0.73423582, -0.38070152, -2.17378081]),
  array([-0.46685237,  1.13995242,  1.72141974]),
  array([-1.05890518,  0.56515737,  0.06632101]),
  array([-0.31354969,  0.18804603, -0.04480358]),
  array([ 0.99426279,  0.84448126, -0.41643312]),
  array([-0.78619416,  0.35856864, -0.15211712]),
  array([ 0.21361536,  1.12046424, -0.04676724]),
  array([-0.37279588,  0.45660945,  0.18525094]),
  array([-0.61233212,  1.21423732, -0.93323259]),
  array([-0.16631633,  0.12469718, -0.25347267]),
  array([ 0.53040565,  0.39292855, -0.30058007]),
  array([-0.27528976,  1.95485232,  0.44248861]),
  array([0.48847648, 1.60586451, 1.64713892]),
  array([-0.72619203,  1.50194201,  0.4755691 ]),
  array([-0.47852288,  0.55683393,  0.18711553]),
  array([1.31635206, 1.80075032, 0.4074105 ]),
  array([-0.0364068 ,  0.92777746, -1.62793745]),
  array([0.03649786, 1.49002466, 1.61838207]),
  array([-0.956195  ,  0.66321265,  0.01630025]),
  array([ 0.49525615,  0.65137334, -2.16610259]),
  array([-0.63503441,  1.26032409, -0.78239321]),
  array([-0.38678471,  0.79612017, -2.31490794]),
  array([ 0.2367994 ,  0.1828797 , -2.46861747]),
  array([-1.12103135,  1.57631566,  0.01648153]),
  array([-0.30977872,  0.35959664, -0.06820283]),
  array([-0.5141658 ,  0.61594439, -0.49538071]),
  array([-0.89829458,  0.66072354, -0.02855191]),
  array([-0.30059286,  1.02558436, -0.29426176]),
  array([-0.56331915, -0.75664411, -2.82161487]),
  array([0.02294657, 0.54128821, 1.12223735]),
  array([-0.38866005,  1.45753926,  0.30832692]),
  array([ 0.11844735,  0.41745606, -0.84994218]),
  array([ 0.34538791,  1.12817968, -0.17471573]),
  array([-0.30211527,  1.5046716 , -0.01105667]),
  array([-0.03186781,  1.06852714, -1.09804689]),
  array([-0.21140005,  0.21673927, -2.04053921]),
  array([-0.1670431 ,  0.10078007, -0.03304415]),
  array([-0.14860507,  0.57212134, -0.2364752 ]),
  array([-0.73843215,  0.01674878, -1.65180738]),
  array([ 0.2831669 ,  0.63337928, -0.46105304]),
  array([-1.05681437,  1.77265831,  0.6996201 ]),
  array([ 0.25533299,  1.06388033, -0.39272548]),
  array([-0.6312178 ,  0.71171884, -0.63572005]),
  array([-0.60234042,  0.75783001, -0.78650217]),
  array([-0.60143952,  1.01218431, -0.97925498]),
  array([-0.53100929,  0.32261395,  0.00905092]),
  array([-0.11116059,  1.33412944,  1.30669824]),
  array([-0.97260886,  0.55369254, -0.07096773]),
  array([0.2745459 , 1.6720336 , 1.47299187]),
  array([ 0.09851475,  2.12309205, -0.41538089]),
  array([-0.54545956,  0.50969284, -0.06503885]),
  array([-0.42752169,  0.47603349, -0.12040932]),
  array([-0.88669652,  0.73219786, -0.37420968]),
  array([ 0.88382562,  0.66697488, -1.08876026]),
  array([-0.72248243,  0.40336019, -0.32569782]),
  array([-0.80879974,  0.67372239, -1.80617003]),
  array([-0.40578149,  1.85671366,  0.51420974]),
  array([ 0.61068808,  0.11651003, -0.85745243]),
  array([-0.60186599,  1.04174592, -0.14707205]),
  array([ 0.31634229,  0.69358041, -0.19986933]),
  array([ 0.09777235, -0.61524014, -2.61971401]),
  array([-0.01239584,  0.57962827,  0.38283904]),
  array([-0.48308864,  0.46386194, -0.00202321]),
  array([-0.30022494,  2.36700612, -1.20723251]),
  array([-0.38791373,  1.41283024, -0.0965381 ]),
  array([-0.94060996,  0.37087275,  0.20995955]),
  array([ 0.33385904,  0.32831318, -0.40629881]),
  array([0.15051675, 0.87603655, 0.38210407]),
  array([-0.34868602,  0.16114437, -0.21617356]),
  array([0.02498706, 0.7927958 , 0.40626377]),
  array([-0.51203008,  1.14923752, -1.70240949]),
  array([-0.00024316,  0.16245338, -0.13362842]),
  array([ 0.74737327,  1.03333885, -1.41977883]),
  array([-0.23029304,  0.74945882, -0.49439196]),
  array([-0.35373161,  0.09987099, -0.81541684]),
  array([-0.36590363,  3.17399849, -2.53843304]),
  array([-0.01403072,  0.08517345, -0.20283568]),
  array([-1.04154461,  1.82617553,  0.1763577 ]),
  array([-1.02654904, -0.00669427, -1.0495018 ]),
  array([-0.81769359,  0.68076051, -0.38520588]),
  array([ 0.38156694,  0.83107213, -2.6003755 ]),
  array([-0.73518808,  0.92502968, -1.52078209]),
  array([ 0.47848469,  0.19623378, -0.22736164]),
  array([-0.60139397,  0.43660494,  0.05577657]),
  array([-0.98845236,  0.2883829 , -0.85245592]),
  array([-0.53951992,  1.34561623, -0.07976222]),
  array([ 0.65565378,  0.50826611, -0.56030908]),
  array([ 0.46822497,  0.68049746, -0.4721584 ]),
  array([-0.35605033,  0.16431468, -1.68625845]),
  array([-0.09154421,  0.74508588, -0.51123171]),
  array([-0.31592579,  0.09110891, -2.7088949 ]),
  array([ 0.14294692,  0.02504607, -0.80968011]),
  array([-0.04473326,  2.0096817 , -1.51721019]),
  array([ 1.37785584,  1.16623039, -2.57876118]),
  array([-0.25029712,  0.70570489,  0.04772869]),
  array([-0.5093321 ,  1.0958843 , -0.10023731]),
  array([-0.74715896,  1.07906489,  1.49090902]),
  array([-0.56288039,  1.53074116, -0.34485918]),
  array([ 0.44939663,  1.13434133, -0.33761679]),
  array([-0.59830024,  1.28540149, -0.43453761]),
  array([-0.2614599 ,  0.48224483,  0.7553208 ]),
  array([-1.01029828,  0.94280135,  0.63828544]),
  array([-0.5469929 ,  1.39303826,  0.04943476]),
  array([-0.44146865,  1.64879813, -2.87465134]),
  array([0.17791307, 1.25062484, 1.02555434]),
  array([-0.43133276,  1.4065102 , -0.87999247]),
  array([ 0.16053398,  0.00232115, -0.8431729 ]),
  array([-0.71207782,  1.1322623 , -0.54183398]),
  array([-0.2497138 ,  1.59268673, -0.22124437]),
  array([ 0.15066053,  0.84561815, -0.17221454]),
  array([ 0.23621831,  0.02072907, -0.24177022]),
  array([ 0.53051028,  0.47365581, -1.07746849]),
  array([-0.37385943,  1.49738675, -0.14886234]),
  array([-0.8023183 ,  0.4226231 ,  0.07104754]),
  array([-0.46538603,  0.12689117, -0.7491274 ]),
  array([-0.722744  ,  1.04113856, -0.92269319]),
  array([0.30824198, 2.39099376, 0.44304334]),
  array([-0.38073887,  0.63232551,  0.56488653]),
  array([-0.19932196,  0.70504068, -0.10334956]),
  array([ 0.52011538,  1.33289676, -0.46371756]),
  array([-0.07033649,  0.89775474, -1.82832403]),
  array([-0.97750637,  1.2059933 ,  0.65425372]),
  array([ 0.50574809,  0.17690078, -1.16274657]),
  array([-0.71807161,  1.36631854,  0.72074967]),
  array([ 0.56890228,  0.16792493, -0.56665846]),
  array([-0.13071101,  1.42955443, -0.14845825]),
  array([ 0.4313616 , -0.48785072, -2.49310616]),
  array([ 0.8233661 , -0.11542119, -2.51817163]),
  array([-0.8511719 ,  0.32950482, -0.95108075]),
  array([0.16760989, 0.85404366, 0.17263003]),
  array([-0.74905676,  1.62329546,  0.18334538]),
  array([-0.26095827,  0.10789294, -0.25355222]),
  array([-0.09201031,  0.3226647 , -0.43503379]),
  array([-0.69302715,  0.92741137,  1.37758019]),
  array([-0.58857366,  0.32350213, -2.59377135]),
  array([-0.03257157,  0.14035429, -1.00156705]),
  array([ 0.01972879,  0.64306248, -0.73737982]),
  array([0.56385422, 1.47451037, 0.10873316]),
  array([-0.19900725,  0.02683513, -0.73431413]),
  array([ 1.15698879,  0.8263254 , -1.1733627 ]),
  array([-0.89971775,  0.35022186, -0.22680217]),
  array([-0.9618578 ,  0.40878881, -0.22036171]),
  array([ 0.82118331,  2.68022365, -0.04925415]),
  array([-0.69890088,  0.25554777, -0.07564012]),
  array([-1.0577729 ,  0.52160289,  0.44043528]),
  array([-0.83994827,  0.87609841, -1.23797983]),
  array([-0.72388985,  0.38205489, -1.94766325]),
  array([-1.09895915,  2.65599105,  1.12950658]),
  array([-0.94337465,  1.37605551,  0.75448506]),
  array([-0.57960811,  0.95870072, -1.11517224]),
  array([-0.94080842, -0.1983283 , -1.49595903]),
  array([-0.55385909,  0.79386647,  0.03913901]),
  array([-0.47453174,  1.89168064,  0.15336063]),
  array([ 0.44865091,  2.47021593, -0.06209171]),
  array([-0.90269853,  1.69456399,  1.06729285]),
  array([-0.26149123,  1.26002703, -0.79387621]),
  array([ 0.67780507,  1.02348173, -0.65528286]),
  array([-0.47664495,  0.68835763, -0.0533873 ]),
  array([ 0.6235495 ,  0.29091399, -1.70490762]),
  array([ 0.58798239, -0.04317778, -1.01068403]),
  array([ 0.3590769 ,  0.12716225, -0.14007411]),
  array([-0.55172883,  1.2960727 ,  0.35274432]),
  array([ 0.67627701,  1.0138529 , -0.22095877]),
  array([-0.80372937,  0.75806138, -0.33403254]),
  array([-0.30942057,  0.29906155, -0.85865502]),
  array([-0.29138382,  2.00496703, -1.35059574]),
  array([-0.6137353 ,  1.90916956, -0.4979173 ]),
  array([0.11288593, 0.29601661, 0.78157576]),
  array([-0.66171233,  1.38380368, -0.28320467]),
  array([-0.65608429,  1.53098216,  0.15790967]),
  array([0.21862918, 1.19433252, 0.6734814 ]),
  array([-0.59613296,  0.78249801, -2.66370259]),
  array([-1.02919382,  0.9238346 , -0.12645624]),
  array([-0.42914441,  0.92845442, -0.11508657]),
  array([-1.20609974,  0.41739058,  0.03972743]),
  array([-0.47358687,  0.68272365, -2.00180542]),
  array([-0.53734263,  0.3742682 , -1.71775841]),
  array([-0.99733597,  0.49373426,  0.40349429]),
  array([-0.07431668,  1.63751003,  0.29572595]),
  array([-0.50905712,  0.9169379 ,  0.50206635]),
  array([-0.65110045,  1.26921697,  1.28188517]),
  array([-0.15246375,  1.094935  , -2.58394904]),
  array([-0.38451764,  1.72132504, -2.38965415]),
  array([ 0.4294503 ,  0.21713128, -0.4371323 ]),
  array([-0.27850749,  1.71180954,  0.18474515]),
  array([-0.45394346,  2.89856545, -0.66430156]),
  array([-0.26502715,  0.74070146, -1.23629858]),
  array([-0.50829211,  0.34842976, -0.2978723 ]),
  array([-0.0957875 ,  2.89888749, -0.31861112]),
  array([ 0.14000298,  0.25744329, -0.93383533]),
  array([-0.33516789,  0.16719914, -2.71642395]),
  array([ 0.02475022,  0.20763562, -0.78929236]),
  array([-0.23631263,  3.41226986,  1.64661914]),
  array([-0.66884231,  2.07736859,  0.14332008]),
  array([-0.0586011 ,  2.32952612, -1.46144102]),
  array([-0.85911313,  0.01852918, -1.65749884]),
  array([-0.76137774,  0.98977215, -0.07333014]),
  array([-0.60363279,  0.68442671, -0.90170242]),
  array([ 0.76044659, -0.00826352, -1.60439968]),
  array([-0.540028  ,  0.6974577 , -2.62903595]),
  array([ 0.13308192,  0.56369852, -0.94136523]),
  array([ 0.73570409,  1.32213603, -0.97302366]),
  array([-0.10490938,  1.09368361, -1.25135022]),
  array([-0.23248826,  0.98688767, -0.84937884]),
  array([-0.7848228 ,  0.6062292 , -0.45463394]),
  array([-0.8501258 ,  0.77295956, -0.01252484]),
  array([-0.69597457,  0.77524084, -0.27363722]),
  array([ 0.00252895,  1.42415111, -0.42913891]),
  array([-0.39728782,  1.50691581,  0.37468451]),
  array([ 0.15302808,  1.21772008, -0.17167376]),
  array([-0.42351648,  2.76938126, -0.35970143]),
  array([-0.80094604,  0.62833797,  0.11600208]),
  array([-1.04717608,  1.79015967,  0.12643628]),
  array([0.43674258, 1.33130796, 1.25278063]),
  array([ 0.86514188,  0.42567912, -3.19494009]),
  array([ 0.33713379, -0.08670248, -1.64337542]),
  array([-1.15934705,  1.7211401 ,  0.00769002]),
  array([-1.14815301,  1.26407145, -0.15191609]),
  array([-0.39337281,  1.01303033,  0.04090757]),
  array([-0.56691538, -0.36354573, -1.80365135]),
  array([-0.97826597,  0.95810102,  0.37924058]),
  array([ 0.04578891,  0.05543382, -1.07136021]),
  array([ 0.75735312,  0.79701889, -0.81597363]),
  array([ 0.51914442,  2.63699437, -0.70295075]),
  array([-0.24984744,  0.59682802,  1.13831704]),
  array([-1.14668785,  1.77908832,  0.70197774]),
  array([-0.0382283 ,  0.7507713 , -0.54621727]),
  array([-0.54542448,  0.00565676, -0.80551165]),
  array([ 0.03251801,  0.30240928, -0.31230041]),
  array([-0.71037786,  0.5717735 ,  0.59481943]),
  array([-0.76353862,  0.28060111, -0.45790001]),
  array([-0.38971165,  1.93544898, -0.88284224]),
  array([0.83094367, 1.03784443, 0.04391736]),
  array([ 0.09381911,  0.4758207 , -0.07757557]),
  array([-0.53385365,  0.24927406, -0.48751859]),
  array([-0.91268965,  1.25384504,  1.70778086]),
  array([-0.24004206,  2.38483119, -0.05528991]),
  array([-0.49450605,  0.89646668,  0.12752502]),
  array([ 0.84741695,  1.04677751, -0.53860429]),
  array([ 0.3270533 ,  0.57197031, -0.34866463]),
  array([0.70862883, 1.75789049, 1.00111223]),
  array([-0.88002003,  1.3928406 , -1.08389472]),
  array([-0.71817069,  1.55876562, -0.83485991]),
  array([ 0.37689344,  0.48903354, -0.17638444]),
  array([-0.25922218,  2.62062252,  1.15135197]),
  array([-1.0352582 , -0.05973057, -1.42617381]),
  array([-0.61837785,  0.68172234, -1.10897763]),
  array([-0.86507766,  2.05679758, -0.32449496]),
  array([-0.1452356 ,  1.65394217,  1.42426837]),
  array([-0.90776115,  1.11940424, -1.14889064]),
  array([-0.79271574,  2.58642311, -0.61332547]),
  array([-1.05005023,  2.49608795,  0.39752124]),
  array([-0.28492123,  2.65134073, -0.22739977]),
  array([-0.20179158,  0.25166264,  0.01976198]),
  array([-0.75025654,  0.47850462,  0.08964268]),
  array([-0.55148588,  0.22251634, -0.08912116]),
  array([-0.70938165,  0.27685107, -0.57344041]),
  array([-0.66131322,  0.39562705,  0.53279382]),
  array([-0.26412519,  1.97052276, -0.13528961]),
  array([ 1.90969013,  0.79688959, -3.09411384]),
  array([ 0.32171339,  1.64294709, -2.91592247]),
  array([-0.39241062,  0.67342172, -0.37081826]),
  array([-0.29395114,  0.70294701,  0.84761531]),
  array([-0.73642263,  0.17633362, -0.35445259]),
  array([-1.0286052 ,  0.60667169,  0.28001638]),
  array([-0.2455775 , -0.07866405, -0.69759091]),
  array([-0.84946692,  0.22071961, -0.31994081]),
  array([-1.12535625,  2.13151848,  1.38158241]),
  array([-0.5947274 ,  1.07037365,  0.08452127]),
  array([ 0.65210758,  1.56078097, -0.11510451]),
  array([ 0.44447954,  1.02912554, -1.18398789]),
  array([0.85868074, 1.97444377, 0.04541353]),
  array([-0.28900022, -0.67451171, -2.48744379]),
  array([-0.17486098,  3.13330558, -0.21247375]),
  array([-0.4018863 ,  0.284896  , -0.88817851]),
  array([ 0.44880757,  0.22787781, -2.05396778]),
  array([-1.58579678e-03,  3.61666515e-01, -2.42848188e+00]),
  array([ 1.10771632,  1.34292047, -1.07834538]),
  array([ 0.38644284,  0.50277996, -0.15735977]),
  array([-0.63546722,  3.04862827,  0.7579935 ]),
  array([-0.55261365, -0.14824349, -1.45794156]),
  array([-1.08406274,  0.5485645 ,  0.29592519]),
  array([ 0.42657556,  0.67447263, -0.39327087]),
  array([-0.13663357,  2.13172598, -1.18251653]),
  array([-0.51000492,  1.12458945,  1.7840356 ]),
  array([ 0.70376233,  0.09440723, -1.52622925]),
  array([-0.59854729,  1.02777757,  1.78298722]),
  array([-0.09399054,  0.36209813, -0.09880606]),
  array([ 0.37367493,  0.71009197, -2.34398047]),
  array([-0.806321  ,  0.94369402,  0.24027642]),
  array([ 1.42792312,  1.55274716, -0.23424849]),
  array([ 1.26500076,  0.28090899, -2.56743983]),
  array([-0.4822186 ,  0.3600655 ,  0.25145351]),
  array([-0.16388703,  0.89492645, -0.17584256]),
  array([-0.19046506,  2.40031807, -0.07594637]),
  array([0.03586086, 0.33348753, 0.25434688]),
  array([ 0.53617658,  0.15279608, -1.3375748 ]),
  array([-0.42644182,  2.01238759, -0.05506204]),
  array([-0.50525333,  0.15570799, -0.42687695]),
  array([-0.79901539,  0.40448756, -0.6049971 ]),
  array([1.14274431, 1.70896057, 0.48129444]),
  array([-0.13763987,  1.44083631,  1.60167485]),
  array([ 0.67549829,  0.27478224, -1.88385248]),
  array([-0.55645381,  1.79247157, -0.29865026]),
  array([ 0.18325805,  2.30664723, -0.02935504]),
  array([-0.78478076,  1.31380261, -0.54980805]),
  array([ 1.03604754,  0.68697876, -2.77131524]),
  array([-0.70795104,  1.83073503,  0.6429874 ]),
  array([-0.16443814,  0.99167751, -1.03852474]),
  array([-0.34721632,  1.70065423, -0.05347976]),
  array([-0.53995964,  0.64490674, -1.53482851]),
  array([ 0.97992961,  0.90111144, -1.05539529]),
  array([-0.57259787,  0.13790575, -0.16574162]),
  array([-0.53565354,  0.99903341,  1.86060899]),
  array([-0.48106083,  1.28312824,  1.1431267 ]),
  array([-0.52430085,  0.83951199,  0.03073456]),
  array([-0.90662677,  0.8310598 ,  1.1460292 ]),
  array([ 0.15891991,  0.21960456, -1.18585481]),
  array([-0.70333159,  1.08213344,  0.10501267]),
  array([-0.33713638,  0.75666408, -0.0460644 ]),
  array([ 0.2956758 ,  0.39639498, -0.22662646]),
  array([-0.17690769,  1.96796848, -0.12487045]),
  array([-0.84364029,  0.34394228, -0.46028952]),
  array([-0.99594943,  0.52691732, -0.8145397 ]),
  array([-0.0166992 ,  1.1011081 ,  1.46936697]),
  array([-0.26681445,  1.09885898, -0.01826603]),
  array([ 0.89000345,  1.65026943, -0.31221306]),
  array([ 0.02994003,  0.87808803, -0.22193065]),
  array([-0.93860885,  1.70430914,  0.58838708]),
  array([-0.57400046,  0.77686316, -0.33769057]),
  array([-0.73180901,  1.6148376 , -1.09599835]),
  array([-0.7009518 ,  1.37566099, -0.69101163]),
  array([ 0.34431548, -0.06235121, -0.3193918 ]),
  array([-0.48308709,  1.18471145, -1.25445649]),
  array([-0.43005122,  0.52968821, -0.06549851]),
  array([-0.99309142, -0.28685162, -1.75618747]),
  array([ 0.81397115,  0.02439318, -2.17962139]),
  array([-0.26707625,  1.16850796, -0.83755352]),
  array([-0.85896045,  1.73120767, -1.43292463]),
  array([ 0.12862867, -0.17197443, -1.04323808]),
  array([-0.36073502,  1.47315883,  1.06898256]),
  array([-0.07291771,  0.35300947, -0.44541743]),
  array([-0.5414652 ,  0.76455476, -1.39057711]),
  array([-0.92935399,  0.54456522, -0.21753821]),
  array([-0.47973897,  0.67087648,  0.16398662]),
  array([ 0.27275564,  2.16549922, -0.34355326]),
  array([-0.97243262,  0.51344324,  0.42263722]),
  array([ 0.50138648,  0.75019558, -0.15316808]),
  array([-0.68798503,  0.99249712, -1.38852277]),
  array([ 0.14245129,  0.7392046 , -0.18224901]),
  array([-0.5409566 ,  1.9163214 , -0.31633377]),
  array([-0.82441572,  2.04971846, -2.26821783]),
  array([-0.19214081,  1.3304913 , -0.31027325]),
  array([-0.91822463,  1.0376895 , -0.11391273]),
  array([-0.25582257,  1.36962151,  0.07847563]),
  array([-0.69312118,  0.58874831,  0.14957318]),
  array([-0.07273757,  0.14189514, -1.16803726]),
  array([0.48174718, 1.19792109, 1.09291687]),
  array([-0.85073328, -0.10347125, -1.37336328]),
  array([-0.92191794,  0.76009825,  1.12178368]),
  array([ 0.00208033,  1.3157695 , -0.41284606]),
  array([-0.55994862,  0.48873263, -0.59167019]),
  array([0.97631981, 1.98837633, 1.22105931]),
  array([ 0.10178878,  1.21462859, -0.84837188]),
  array([ 1.34381043,  0.88219685, -2.81394129]),
  array([0.39188992, 0.73529743, 0.17637541]),
  array([ 0.52756412,  1.44321117, -1.21115582]),
  array([-0.26987378,  0.62730773,  0.32787333]),
  array([-0.39548968,  1.53211456, -1.02389008]),
  array([-0.02738891,  0.41865109, -0.16054899]),
  array([0.69669321, 2.05430907, 0.14502766]),
  array([-0.34049588, -0.11145625, -0.95956469]),
  array([-0.33446556,  0.92910088, -0.42917552]),
  array([-0.41959838,  0.77962238, -0.25972898]),
  array([0.45785522, 0.84513589, 0.26861174]),
  array([-0.19597258,  1.69182726,  0.66765464]),
  array([ 0.72423426,  0.42797285, -1.15255246]),
  array([ 0.70135224,  0.31810273, -1.27522076]),
  array([-1.07636685,  0.64457617,  0.80368657]),
  array([-0.07185859,  1.57836396, -0.17321507]),
  array([-0.34275115,  0.72183523, -0.2377774 ]),
  array([-0.55409361,  0.88133942, -0.81155875]),
  array([ 0.01185501,  1.36747   , -0.11688775]),
  array([-0.25950818,  0.36560058, -0.1402499 ]),
  array([-0.00954233,  1.90087444,  1.67375361]),
  array([-0.8879016 ,  1.95104961,  1.06964886]),
  array([ 0.15271864,  0.57711793, -0.8989272 ]),
  array([-0.6507134 ,  0.57780749,  0.61143081]),
  array([-0.60712193,  0.62074653, -0.81023111]),
  array([ 0.76266112,  0.08498727, -1.15979061]),
  array([-0.09709626,  1.2554866 ,  1.79118934]),
  array([-0.1959716 ,  0.09893254, -0.80437219]),
  array([ 0.21576663,  0.58055574, -1.28201993]),
  array([-1.16316073,  0.54075368,  0.13690175]),
  array([-0.05750604,  1.08884749, -0.11854578]),
  array([-0.8656069 ,  0.42805644, -0.57673266]),
  array([-0.67237044,  1.29734747, -0.02420279]),
  array([-0.48118721,  1.03396502,  0.29080433]),
  array([ 1.14178765,  2.1258892 , -2.19227945]),
  array([-0.80731038,  0.96522534, -0.00169399]),
  array([-0.2145914 ,  0.15784505, -1.3728619 ]),
  array([ 0.3522276 ,  2.74608135, -0.34958571]),
  array([-0.08178023,  1.29010796,  1.40465237]),
  array([ 0.85831407,  1.14117572, -0.12650189]),
  array([-0.68658809,  0.0976999 , -0.59935819]),
  array([-0.28847014,  0.19395288,  0.13009338]),
  array([-0.51973363,  0.23825741, -0.7897186 ]),
  array([-0.11921834,  1.58972986, -0.50846341]),
  array([-0.92232509,  1.36857418,  0.20351485]),
  array([-0.94362934,  1.21548972,  0.32551576]),
  array([ 0.04236669,  0.19867855, -2.46296265]),
  array([-0.30328288, -0.14848791, -0.96879929]),
  array([-0.5468534 ,  1.94986391,  1.35716057]),
  array([0.21612442, 1.15243705, 0.18835253]),
  array([ 0.81406261,  1.35232277, -1.05875654]),
  array([-0.63554866,  0.87409767, -0.42178441]),
  array([-0.13846115,  0.77023397,  0.3574999 ]),
  array([-0.32917403,  0.77411763, -0.04002602]),
  array([-0.41902128,  1.47206696, -1.78913537]),
  array([-0.49871096,  1.03614706, -0.34658253]),
  array([-0.48674407,  0.53223373, -0.44683397]),
  array([-0.77305659,  1.06499113,  0.13939872]),
  array([ 0.15832805,  1.60962472, -0.06193776]),
  array([-0.65555124,  0.3219389 ,  0.03100985]),
  array([-0.56097004, -0.1234733 , -1.99066424]),
  array([ 0.75455746,  0.75810468, -0.34295611]),
  array([-0.35728023,  1.25599651, -0.81528409]),
  array([ 0.02518083,  0.79592378, -1.59314832]),
  array([-0.78992437,  0.63568824,  0.36100123]),
  array([-0.41609769,  0.81251469, -0.09576119]),
  array([-0.69270266, -0.01037937, -1.85210736]),
  array([ 0.02616861,  0.64765324, -0.09193155]),
  array([ 0.53512928,  0.15419283, -0.48098864]),
  array([-1.10159568,  0.97434467, -0.0104941 ]),
  array([-0.90924468,  1.02594467,  1.17080809]),
  array([-0.04688348,  0.59532308, -0.15878534]),
  array([-0.28133293,  0.22674042, -1.02581859]),
  array([-0.94720748,  2.05647501,  0.25719234]),
  array([ 0.25954134,  0.51562882, -0.16006301]),
  array([-1.02961748,  0.97957407,  0.69200262]),
  array([ 0.03540652,  0.05028964, -2.53312214]),
  array([-0.07940358, -0.27955163, -1.56200163]),
  array([-0.75351075,  1.29412378, -0.19650257]),
  array([ 0.91539308,  1.61422515, -0.16625811]),
  array([-1.04283724,  1.24636982,  0.36181024]),
  array([0.03166406, 1.18847521, 0.66826852]),
  array([ 0.03945723,  1.28780963, -0.24251681]),
  array([-0.48641559, -0.05609449, -0.75387739]),
  array([-0.32207038,  0.29570509, -0.02895257]),
  array([ 0.53569258, -0.96789606, -3.00879572]),
  array([-0.20902247,  0.66059864, -0.30831414]),
  array([ 0.35199508,  1.29112524, -0.30416849]),
  array([-0.68859128,  0.37579978,  0.07210794]),
  array([ 2.05357031,  1.92812682, -1.0591993 ]),
  array([-0.68175292,  2.81541833,  1.01391683]),
  array([-0.48584076,  0.46150921, -1.11101755]),
  array([-0.96598988,  0.71931205,  0.28020649]),
  array([-0.35404439,  1.49720283,  0.27353026]),
  array([-0.30550051,  1.36667328,  0.00345022]),
  array([-0.67828043,  0.54085141, -0.40933692]),
  array([-0.68799376,  1.42764486, -0.34068435]),
  array([-0.69433396,  0.11627884, -1.45623189]),
  array([ 0.01967933,  0.99991092, -1.72128072]),
  array([ 1.56560543,  1.69479931, -0.71316774]),
  array([-0.18391521,  0.30979555,  0.41575639]),
  array([-0.46449802,  1.05828576,  0.4420038 ]),
  array([-0.54371837, -0.41435881, -2.91501845]),
  array([ 0.39322415,  0.67687263, -1.07513523]),
  array([-0.28936683,  1.25264752, -0.27425947]),
  array([0.02621983, 0.72027452, 1.64595307]),
  array([-0.86847014,  1.46671072,  2.00173711]),
  array([-0.43052569, -0.0748349 , -0.80416086]),
  array([-0.14923012,  1.21665153, -0.77237078]),
  array([-0.59204051,  1.29528994,  0.1336048 ]),
  array([-0.48808996,  0.06963643, -0.54248179]),
  array([-0.52357825,  1.56492744, -1.25034484]),
  array([-0.44155843,  0.95439167, -0.60962786]),
  array([-0.96285034,  0.08696776, -0.71626913]),
  array([-0.34199783,  2.27292233, -1.24975859]),
  array([0.88038942, 1.3989835 , 0.63620073]),
  array([-0.37985636,  1.99536874, -0.72570025]),
  array([-0.09248009,  1.91833042, -0.27933451]),
  array([-1.09890093,  0.83638959,  0.05454605]),
  array([-1.07137659,  0.60498853,  0.01656574]),
  array([0.07363891, 2.15189615, 1.2030434 ]),
  array([-0.65023205,  0.10246879, -0.39311484]),
  array([-0.05359209,  0.83576635, -1.03601889]),
  array([-1.18872207,  0.69510303,  0.36881968]),
  array([-0.33209893,  0.68185087, -0.96457827]),
  array([ 0.28513689,  0.35056315, -0.29651803]),
  array([ 0.71591599, -0.17687491, -1.84807916]),
  array([-0.58532329,  0.74291952,  0.38312875]),
  array([-0.3405089 ,  1.67608198, -0.23146683]),
  array([-1.13065652,  0.58749683,  0.29190852]),
  array([-1.14693064,  1.80636558,  0.62671217]),
  array([-0.86672788,  0.88521211, -1.49025557]),
  array([ 0.67405071,  1.31353663, -0.05700422]),
  array([ 0.03400316,  1.395709  , -0.30037892]),
  array([ 0.3950331 ,  0.55742763, -0.29563952]),
  ...],
 [array([-0.87801333, -0.76353628, -0.30746848]),
  array([-0.33529904, -0.02377375, -0.20003663]),
  array([-0.38449538, -1.31623257,  0.84276872]),
  array([-0.91836719, -1.03324704, -0.74590805]),
  array([-1.19278484, -1.10303746,  1.84039163]),
  array([-0.22698829,  0.47204524,  1.59431983]),
  array([-1.08695117, -0.9339755 , -0.27501231]),
  array([-0.23178705, -0.6814469 , -0.68100015]),
  array([-1.09629559, -0.8411248 , -0.05262142]),
  array([-1.03740551, -1.31524731, -0.03756151]),
  array([-1.12239667, -0.34913811,  0.28316616]),
  array([-0.90716762,  0.62061848,  1.76138683]),
  array([-0.5266677 , -1.48630234, -1.1187748 ]),
  array([-0.91565249,  0.03391659, -0.20666188]),
  array([-1.26408687, -1.29502596,  1.08227509]),
  array([-0.74835104, -0.80660014, -0.49787758]),
  array([-1.01033507, -0.14757181, -1.09837537]),
  array([-0.5410388 , -0.7801968 ,  0.42932549]),
  array([-0.87527187, -0.77398992, -2.30972787]),
  array([-0.5273802 ,  0.22401905,  1.17645536]),
  array([-0.9788208 , -1.2432527 ,  0.14309746]),
  array([-1.17578698, -0.07064457,  0.71979437]),
  array([-0.81378235, -1.00343332, -1.15238736]),
  array([ 0.23422553, -1.31260695, -1.18902584]),
  array([-0.16207183, -0.56407048, -0.15149845]),
  array([-1.06366883, -1.77996929, -0.65914903]),
  array([-1.03180198, -0.95880744,  0.03722377]),
  array([-0.46085831, -0.63973412, -1.76267227]),
  array([-0.66977443, -1.00630598, -0.6180389 ]),
  array([-0.91252999, -1.24876457, -0.37019571]),
  array([-0.66295498,  0.51715473,  1.63756463]),
  array([-0.79032092, -1.5699541 , -2.71457383]),
  array([-0.63913439, -1.34876645,  0.02325512]),
  array([-0.94635554, -0.06989358,  1.46667131]),
  array([-0.80850848, -1.27627547, -1.33921874]),
  array([-0.2865632 , -0.18762082, -0.31314995]),
  array([0.12861946, 0.09165273, 1.783508  ]),
  array([-0.86518771, -0.60721551,  0.06373636]),
  array([-0.73413486, -0.32022185,  1.52727591]),
  array([-0.93160505, -0.86649674, -0.27712906]),
  array([-1.08877987, -0.12896975,  0.03387868]),
  array([-0.82594655, -1.47372744, -0.01277726]),
  array([-0.84937533, -1.83664417, -2.65542209]),
  array([-0.39440513, -0.20718334,  0.72892054]),
  array([-0.90775021, -1.57206686, -1.89753477]),
  array([-0.88570544, -0.42318895,  0.15539842]),
  array([-1.02389297, -0.86027334, -0.7270954 ]),
  array([-0.87426621, -0.96410838, -0.0543772 ]),
  array([-0.94142054, -0.83787736, -0.22962382]),
  array([-0.59383664, -0.88866989, -1.24577316]),
  array([-0.56777486, -0.0088887 ,  0.78935739]),
  array([-0.61074334, -0.04511405, -0.38941347]),
  array([-1.03556331, -1.24999874,  0.13850973]),
  array([-0.58941818, -0.99685484, -0.19000358]),
  array([ 0.28139727, -0.44620146, -0.25244201]),
  array([-1.14240811, -1.25984546,  1.30202691]),
  array([-0.09355925, -1.46152099,  1.02515578]),
  array([-0.96074903, -0.53966421, -0.28015962]),
  array([ 0.21062569, -0.17558406,  0.67383341]),
  array([-0.7955157 ,  0.21740656,  0.28503086]),
  array([-0.91851731,  0.21545086,  1.51188159]),
  array([-0.69449489, -0.90039607,  0.57819559]),
  array([-0.08731566, -1.11344255, -0.1019749 ]),
  array([ 0.05244876, -0.33630598, -0.40263296]),
  array([-0.18300947, -0.52530092, -1.25885672]),
  array([-1.04904411, -1.13822783,  0.21400879]),
  array([-0.94281094, -1.09135356, -0.6236314 ]),
  array([-0.41862427, -0.81690817,  0.22546605]),
  array([-1.10202022e+00, -8.39610785e-01,  7.07011768e-04]),
  array([-0.63604632,  0.19436615,  0.13555537]),
  array([-1.32943543, -1.27774302,  1.69103637]),
  array([-0.50686532, -0.4690193 , -0.7739866 ]),
  array([-0.67340945,  0.19768371,  1.05055923]),
  array([-0.95434198, -0.07852824,  0.21290056]),
  array([ 0.21641926, -0.34217432, -0.15826699]),
  array([-0.72146654, -0.19973744, -0.16075033]),
  array([ 0.39438079, -1.31102984, -0.27975718]),
  array([-0.68470939,  0.12039409, -0.07864354]),
  array([-0.5968026 , -1.05101625, -0.63899192]),
  array([-0.62047824, -1.1035217 ,  1.96628489]),
  array([-0.13768862,  0.08748086,  0.52276118]),
  array([-0.83807248, -0.26636109, -0.18696782]),
  array([ 0.23198701, -0.73300073,  0.85303662]),
  array([-0.99054223, -1.76613966, -1.06348189]),
  array([-0.29454802, -0.91966404, -0.50019702]),
  array([-0.71745287, -1.70317127, -0.54045234]),
  array([-0.77717326, -0.11634218, -0.86295617]),
  array([ 0.29300809, -0.59747504, -0.26802379]),
  array([-0.87296371, -0.5364729 , -0.35164151]),
  array([-0.55034485,  0.419394  ,  0.75947402]),
  array([-1.06829045, -0.25466745,  0.03764964]),
  array([-0.55417532, -1.13245826, -1.43812038]),
  array([ 0.09319078, -1.2656755 , -0.56443512]),
  array([-0.68387904, -1.72964128, -0.79088752]),
  array([-1.0772315 , -1.49407395, -0.0418289 ]),
  array([-1.10808787, -1.68643618,  0.64758033]),
  array([-1.38152883, -1.26396571,  2.17631777]),
  array([-1.09727536, -1.5551122 , -0.40887927]),
  array([-1.12768378, -1.51390033,  0.15127901]),
  array([-0.46091836, -0.74741676, -0.8153683 ]),
  array([ 0.02547287, -0.40265634, -0.1320199 ]),
  array([-1.02955049, -1.02853612, -0.84910834]),
  array([-0.69713555,  0.15313945, -0.0406216 ]),
  array([ 0.48464996, -1.74595814, -0.31254198]),
  array([-1.04612528,  0.01356933,  0.18258991]),
  array([-0.85965128,  0.10951414,  1.18846355]),
  array([-0.99625475, -0.20854363,  0.02002228]),
  array([-1.07212366, -0.46029246,  0.94681183]),
  array([-1.18916639, -1.74677856,  0.5099364 ]),
  array([-0.08232791, -1.34849807,  0.89462985]),
  array([-0.68404875, -1.29534496, -0.01510343]),
  array([0.18265807, 0.24542121, 0.99757583]),
  array([-1.16348774, -0.56117517,  0.27300722]),
  array([-0.39403969, -1.05090877, -0.07551617]),
  array([-0.51268501, -0.51281195, -0.44164378]),
  array([-0.84126255, -0.08672795, -0.06260532]),
  array([-0.05119352,  0.24498026,  0.97720756]),
  array([0.03471359, 0.32089714, 0.90309291]),
  array([-0.73682096,  0.00106692,  0.99112803]),
  array([-1.08165674,  0.05204956, -0.16328695]),
  array([-0.59582173, -0.62110441,  1.11122218]),
  array([-0.36216422, -0.34477379,  0.02052853]),
  array([-0.0391472 ,  0.35939284,  1.12833608]),
  array([-1.15258348, -1.44097572,  0.93700569]),
  array([-0.75763073, -0.7045789 ,  0.09507979]),
  array([-0.26472556, -0.50360673, -0.73420674]),
  array([-0.77648322, -0.98585024, -1.92723948]),
  array([-0.90827986, -0.05599592,  0.30753241]),
  array([-0.03297601, -0.64326808, -0.95890829]),
  array([-1.26706479, -1.04509965,  1.32410643]),
  array([-0.30505424, -0.42442603,  0.33213014]),
  array([0.02772371, 0.36939363, 1.8604448 ]),
  array([-0.79168783, -1.36615902,  0.07591749]),
  array([-0.89214369, -1.14425622, -2.16850949]),
  array([-0.70923437, -0.76399993, -2.42776078]),
  array([-0.79305402, -0.93318583, -0.14001361]),
  array([-0.9418794 , -1.70825472,  0.1534103 ]),
  array([-0.719391  , -1.43635416, -0.12293285]),
  array([-0.41368242, -1.41355087, -2.86733874]),
  array([-1.01916703, -1.47626079, -0.30584892]),
  array([-1.16370002, -0.26271506,  1.81820909]),
  array([-0.73336046, -1.14129076, -1.49426768]),
  array([-1.23978315, -1.73339178,  0.98146202]),
  array([-1.24425107, -1.154553  ,  1.49063774]),
  array([-0.63882021, -0.91413714, -1.86374987]),
  array([-0.47560749, -0.56104671, -1.08048383]),
  array([-1.01537522,  0.1701125 ,  0.44740357]),
  array([-0.48209648, -1.78286103, -1.54509259]),
  array([-1.14010987, -1.51061396,  0.26703553]),
  array([-0.61405044, -1.21172732, -1.18139617]),
  array([-0.85207731, -1.67324414,  1.48480634]),
  array([-0.34821752, -0.92993893, -0.14256227]),
  array([-1.26344584,  0.21658801,  0.63673626]),
  array([-0.15202123, -1.68222536, -0.78639826]),
  array([-0.27474058, -0.25183637, -0.4147247 ]),
  array([-1.01248641, -0.55080666,  0.20074839]),
  array([ 0.03313747, -0.75300262, -0.16084278]),
  array([-0.6204836 , -0.49614127, -0.01066398]),
  array([-1.13890161, -1.76007224,  0.04168969]),
  array([-0.22302522, -0.12604925, -0.01970088]),
  array([-1.18812362, -1.3151162 ,  0.37463167]),
  array([-1.13471906, -0.61499375,  0.24252107]),
  array([-0.86083912, -0.51837151,  0.52010955]),
  array([-0.83250823, -0.7485732 , -0.00595597]),
  array([-1.10040909, -1.05600596,  0.04849366]),
  array([-0.97517875, -0.47656995,  2.03388611]),
  array([-0.7246048 , -0.63345933, -0.90505774]),
  array([-1.21504397, -0.4441202 ,  0.37422962]),
  array([-1.23298596, -1.23691201,  1.34632242]),
  array([-0.11858371, -1.17528018, -0.02305733]),
  array([-1.10369173,  0.18151719,  0.17322747]),
  array([-0.76091457,  0.02739253,  0.2662168 ]),
  array([-0.97122045, -0.56376884, -0.19146569]),
  array([-0.9903694 , -0.03713642, -0.57010964]),
  array([-0.4014754 , -1.03290309, -0.65414604]),
  array([-0.95920159, -1.27615097, -0.92725843]),
  array([-0.60549463, -0.52090256,  0.10462709]),
  array([-1.11636015, -1.51689513,  0.04579258]),
  array([-0.46281729,  0.14041863,  1.64306586]),
  array([-0.92914828, -1.81554637, -1.91228904]),
  array([ 0.11315449, -1.69265776,  0.23517027]),
  array([-0.009205  , -0.84022178,  0.01350384]),
  array([-0.51880821,  0.18543622,  0.72474177]),
  array([-0.51579378, -0.85441954,  0.15204095]),
  array([ 0.27778356, -0.76366167,  1.87480607]),
  array([-1.14253754, -1.44363261,  0.84342171]),
  array([-0.48540204, -0.04829893, -0.29734209]),
  array([-0.95962537, -0.51831566,  0.72586302]),
  array([ 0.26504777, -0.49910234,  0.72802478]),
  array([-0.0428217 , -0.67255602, -0.96336173]),
  array([-0.72056395, -0.10000306, -0.27301635]),
  array([-1.14191041, -1.14593983,  1.37241099]),
  array([-0.75499965, -0.78313672,  0.2126123 ]),
  array([-0.93307529, -1.25788795,  0.2814771 ]),
  array([-0.19180448, -0.89922878, -1.60180852]),
  array([-1.10520845, -1.73581352,  0.00470311]),
  array([-0.86123377, -0.83174267,  0.308389  ]),
  array([-0.85582762,  0.19886826, -0.12970819]),
  array([ 0.24345467, -0.75203913, -0.33082054]),
  array([ 0.18098631, -1.66539868,  0.78004841]),
  array([-0.53763267,  0.52535133,  1.57074663]),
  array([-0.7314177 , -1.0802764 , -1.68120794]),
  array([-0.77748074,  0.13578799, -0.3088789 ]),
  array([-1.00495918, -0.98635214,  0.2879648 ]),
  array([-0.39065771, -0.04974095,  1.32577663]),
  array([-0.83493711, -1.45660144,  0.13069291]),
  array([ 0.58476692, -1.73686272,  0.54967443]),
  array([-1.04609593, -1.04836405,  0.59665328]),
  array([-1.02217002, -1.35900647, -1.17132609]),
  array([-1.08960772, -0.27623609,  0.25666354]),
  array([ 0.28320735, -1.49617245, -0.03802429]),
  array([-0.54750486, -0.56029244,  0.70695171]),
  array([-0.81057576, -1.04671928, -1.89715311]),
  array([-1.0743712 , -1.74396915, -0.28256446]),
  array([-0.80745758, -0.47753795, -0.02395451]),
  array([-0.29548878, -0.55235552, -0.13384266]),
  array([-0.27493189, -0.91684425, -0.08153235]),
  array([ 0.34237013, -0.71168891,  0.43711413]),
  array([-0.67429587,  0.17655624,  0.06667635]),
  array([-1.00883826,  0.02931516, -0.20636207]),
  array([-0.88242117, -0.50803413,  0.24906415]),
  array([-0.25815347, -0.25132868, -0.95060796]),
  array([-1.0413714 , -1.2454164 , -0.11342091]),
  array([-1.01472946, -0.8867369 , -0.39403006]),
  array([-0.6458849 ,  0.40328116,  1.32771791]),
  array([-0.57832809, -1.04041121,  0.5901146 ]),
  array([-0.87516856, -0.46300981, -1.48542784]),
  array([-1.05208479,  0.13383762, -0.19498767]),
  array([-0.48091573, -0.08366296, -0.63917119]),
  array([-0.78361914, -0.07156274,  0.07412664]),
  array([-0.96639623, -1.45048747,  0.19439613]),
  array([0.13457567, 0.38378319, 1.61303004]),
  array([-0.42400636, -0.57025806, -1.90344435]),
  array([-0.94832449, -0.49081985, -0.03952321]),
  array([-0.93697178, -1.12975104,  0.1737538 ]),
  array([-1.38618663,  0.28221931,  2.0570262 ]),
  array([-0.79615085, -0.81551384,  0.43762823]),
  array([-0.85494826, -1.65703869,  1.85071913]),
  array([-1.05834431, -1.46589945,  0.05911105]),
  array([-0.98945952, -1.48309567,  0.21689954]),
  array([ 0.27146314, -0.72163868, -0.46491546]),
  array([-0.28618051, -0.22197754,  0.03384475]),
  array([-1.01324209, -0.19949188,  1.51370725]),
  array([-1.0175538 , -1.32705768, -0.93744348]),
  array([-0.82236366,  0.24386502,  0.13048863]),
  array([-0.88156616, -0.97347418, -0.13696247]),
  array([-1.14248839, -1.47681526,  0.56607833]),
  array([-0.4007499 , -0.39265726,  0.01527346]),
  array([-0.84386956,  0.41515261,  0.54511603]),
  array([-1.35483064, -1.70296481,  2.05319873]),
  array([-0.85177788,  0.55980149,  1.15951003]),
  array([-0.86481796, -0.15444742,  1.86532343]),
  array([-0.81305784,  0.08246111,  0.09118675]),
  array([-0.94859401, -1.15184633, -0.38868177]),
  array([-0.08961349, -0.21132644,  0.94982873]),
  array([-0.65785508, -0.21139448,  0.43541693]),
  array([-1.06867607, -1.69369311,  0.67843394]),
  array([-0.44200018, -1.23353178, -1.38877677]),
  array([-1.12090627, -1.29972373,  0.02534706]),
  array([-0.57442225, -0.53171131,  0.0019544 ]),
  array([-0.83516142, -0.13304313,  0.30072153]),
  array([-1.0479453 , -0.36763218,  0.1430122 ]),
  array([-0.21834563, -0.52802318,  1.56900447]),
  array([-1.06040985, -0.82603985,  0.64756399]),
  array([-0.66777643, -1.01433867, -1.51524721]),
  array([-0.64643902,  0.17242519,  0.05706087]),
  array([-0.83921812, -1.23151925, -0.64333771]),
  array([ 0.01730109, -0.09668187,  1.64992524]),
  array([-0.84823976, -1.44540059, -0.23141879]),
  array([-1.12524088, -1.06716211,  0.3420994 ]),
  array([-0.08313444, -1.64852136,  1.34577544]),
  array([-0.60297262, -1.53473838, -0.56026862]),
  array([ 0.16656423, -0.43738048, -0.60717356]),
  array([-0.52857862, -0.64628136, -0.80197861]),
  array([-0.95563559, -0.69552617, -1.70263489]),
  array([-0.895109  , -0.71643868,  1.39923599]),
  array([-0.72256447, -1.20430436,  1.05532805]),
  array([-0.61602389,  0.20977606,  1.35633208]),
  array([-0.41789562,  0.06400108, -0.08582778]),
  array([0.18436138, 0.01819071, 1.34704978]),
  array([-1.22466232, -1.73739084,  0.84060235]),
  array([-0.65256843, -0.88474066, -0.02774763]),
  array([-0.63220958, -1.11379081, -0.97494238]),
  array([-0.78858302, -0.36691481,  0.35793144]),
  array([-1.19682445,  0.4825732 ,  1.50443837]),
  array([-0.88621837, -0.9311652 , -0.04546331]),
  array([-0.98066521, -0.60952682, -1.43009487]),
  array([-0.45788656, -0.47202212,  0.88255082]),
  array([-0.66679995, -0.40864592, -0.47993172]),
  array([-0.47608477, -0.1354513 ,  0.25312271]),
  array([-0.15098777, -1.17323561, -1.15511314]),
  array([-0.96901468, -0.96158791,  2.03965838]),
  array([-0.82604173, -0.31889501, -1.39314035]),
  array([-0.79170228, -1.49995633,  0.16949527]),
  array([-0.73365729, -0.19141573, -0.07922765]),
  array([-0.62913785, -1.07457865, -0.17276402]),
  array([-0.94210577, -0.64964154,  0.21901647]),
  array([-0.60377943, -0.03403252, -0.29426293]),
  array([-1.18005763, -1.4668792 ,  0.91605847]),
  array([-1.13092955, -0.81436684,  1.9738326 ]),
  array([-0.49080632,  0.16114657,  1.68802632]),
  array([-1.04720731, -0.32189923,  0.33916896]),
  array([-0.59215151, -1.7346363 , -2.46969937]),
  array([ 0.27905379, -1.71461172, -0.25090099]),
  array([-0.00380492, -0.05882153,  0.71290111]),
  array([-0.94603059, -1.7779118 , -1.47813463]),
  array([-0.8482518 , -1.12758236,  0.0326569 ]),
  array([-0.06310307, -0.49100345,  0.37903342]),
  array([ 0.18651522, -0.99756931, -0.69427933]),
  array([-0.87004724, -1.06137638, -0.4752254 ]),
  array([-0.68427801, -1.02086611, -1.65111197]),
  array([-1.04897593, -1.71751587, -0.24225126]),
  array([-0.48004258, -0.07245067, -0.16463851]),
  array([-0.71765857, -1.00108605, -0.36501982]),
  array([-1.11323591, -1.73369048,  0.0794837 ]),
  array([0.06824668, 0.17311783, 1.7324413 ]),
  array([-0.56035849, -0.66162645,  1.88673244]),
  array([ 0.0071644 , -1.71555537, -0.90882941]),
  array([-0.92052539, -0.55345702,  0.16844335]),
  array([ 0.27790192, -1.22081505, -0.15462218]),
  array([-0.98190525,  0.08877163, -0.58335966]),
  array([-0.66004552,  0.15485648,  1.04809117]),
  array([-1.04285408, -0.84840637,  0.54630406]),
  array([ 0.19759716, -0.36637694, -0.33212675]),
  array([-0.46164216, -0.18508735, -0.0398176 ]),
  array([-0.98581384, -1.45191216, -0.33966842]),
  array([-0.58181668, -0.43947128,  0.05509084]),
  array([-0.70354386, -0.89545421,  0.02499536]),
  array([-0.92645212, -1.46105162, -0.17770737]),
  array([-0.66736617, -1.48691369, -2.3296082 ]),
  array([-0.33768543, -1.6693989 , -0.51583665]),
  array([-0.37448706, -0.38933512, -0.18763733]),
  array([-0.94224725, -1.41714445,  0.96127061]),
  array([-1.18727226, -0.15310702,  1.30620174]),
  array([-1.06576295, -1.49394063,  0.24933303]),
  array([ 0.21351613, -0.99415363, -0.34040652]),
  array([-1.04407057,  0.23709323,  0.42151173]),
  array([-1.08384179, -0.09710609,  0.26476294]),
  array([-0.94282169, -0.81118074,  0.42549122]),
  array([-0.85072247, -1.51424973, -1.16006022]),
  array([-0.81450074, -1.06147392, -0.41190525]),
  array([-0.07380947, -0.6861286 ,  1.52014975]),
  array([-0.74546215, -0.48573606, -0.34220198]),
  array([-1.12085505, -1.26656762,  0.3017554 ]),
  array([-0.64671492, -1.00042105, -1.6805637 ]),
  array([-0.21594621, -1.25666481,  0.61162238]),
  array([-1.19278916, -0.450006  ,  0.16691272]),
  array([-0.48837802, -0.46433856, -0.01280754]),
  array([-1.1505386 , -1.21250709,  0.34076415]),
  array([-0.21544948, -0.7830851 , -0.23946349]),
  array([-1.3151632 , -1.71345578,  1.68367259]),
  array([-1.17641856, -1.75015002,  0.39118264]),
  array([-1.10680245, -1.20394485,  0.72461803]),
  array([-1.06366883, -1.77996929, -0.65914903]),
  array([-0.59728076, -1.68438086,  1.16326232]),
  array([-0.2262692 , -0.03440526,  0.68253197]),
  array([-0.91742651, -0.85720944, -0.11753365]),
  array([-0.88184439, -0.83048012, -0.69710078]),
  array([-1.28450695,  0.44725678,  1.46545226]),
  array([ 0.15460479, -0.70629105,  0.47032859]),
  array([-1.01247152, -0.63083224,  1.6405396 ]),
  array([-0.94516733, -1.03976641,  0.09731265]),
  array([-0.33719681, -0.09516964, -0.36437684]),
  array([-0.62223111, -0.95134132, -0.47829212]),
  array([-0.84804503, -1.12070365, -0.73830544]),
  array([-0.51576458, -1.05376541,  0.03839156]),
  array([-0.15338686, -0.71260936,  0.76836435]),
  array([-0.77630502, -0.78179098,  1.21407147]),
  array([-0.76073139, -0.13323509, -0.65654869]),
  array([-0.05887124,  0.34065747,  1.27055662]),
  array([-1.01531736, -0.57983392,  0.2671878 ]),
  array([-1.11954444, -1.26691424,  0.28954632]),
  array([-1.05532117, -0.4888259 ,  0.50354273]),
  array([-0.5930199 , -0.54175081, -1.66008814]),
  array([-0.07502239,  0.04525061,  0.01571291]),
  array([-0.87219301, -1.41460056, -0.04485218]),
  array([-1.02478465, -0.31177608,  0.89614173]),
  array([-0.83904877,  0.71488394,  1.9880645 ]),
  array([-0.67068758, -0.59880276,  0.17070464]),
  array([-1.16056652, -1.75434246,  0.24351132]),
  array([-0.16023697, -0.42573677,  1.46017922]),
  array([-1.17077945, -1.17372298,  2.06126392]),
  array([-0.93747845, -0.07582899, -0.62331751]),
  array([-0.604359  ,  0.16362375,  0.01043691]),
  array([-0.16377903, -0.78005081, -0.52563413]),
  array([-1.02122928, -0.52332981,  0.1968694 ]),
  array([-1.31078099, -1.71461476,  1.64284966]),
  array([-0.88142999, -0.19205477,  0.13985581]),
  array([-0.75638135, -0.3144353 ,  1.82708984]),
  array([-1.19008399, -1.06229257,  1.00498223]),
  array([-0.82551987, -0.34901903,  0.5035947 ]),
  array([-1.27151586, -1.50903026,  1.21427602]),
  array([-1.04333144,  0.16041949,  0.15036218]),
  array([-1.2355069 , -0.39670119,  1.3095613 ]),
  array([-0.91458696, -0.08424669,  0.23423001]),
  array([-0.90740589, -0.19528674, -0.40525454]),
  array([ 0.14378361, -0.43650063, -0.07105992]),
  array([-1.08961076, -0.66009334, -0.45457546]),
  array([-0.95651773, -1.45863824,  0.18691237]),
  array([-0.8476834 , -0.21559452, -0.66832692]),
  array([-0.89221612, -0.88015753,  0.2199428 ]),
  array([-0.45511167, -1.22659794,  0.46755213]),
  array([-0.10605418, -0.68219846,  1.08846361]),
  array([-1.17305161, -1.53507141,  0.29702217]),
  array([-1.10555619, -0.75929593,  0.07302075]),
  array([-1.18723285, -1.09938271,  0.30353831]),
  array([-0.9717028 , -1.09814089,  1.99036692]),
  array([-0.96737053, -0.58974719, -0.36466196]),
  array([-0.94238128, -1.71714087,  0.53382615]),
  array([-0.79606158, -1.53526619, -2.3842106 ]),
  array([-0.0413156 , -1.00589632, -1.10254234]),
  array([-0.98188176, -0.06554032, -0.44308568]),
  array([-0.26135096,  0.09517648,  0.22680029]),
  array([-0.9312772 , -1.43316599, -0.5708246 ]),
  array([-0.87223463, -0.58187567, -0.11162355]),
  array([-1.18430197, -0.65458393,  1.88571628]),
  array([-1.24516705, -1.15431075,  1.49917073]),
  array([-0.67478155, -0.63134302,  0.15013035]),
  array([-0.92045418, -0.85557733,  0.1074939 ]),
  array([-1.13911933, -1.07893779,  0.13221733]),
  array([-1.1539813 , -1.75608407,  0.18216604]),
  array([-0.79804466, -0.29196906, -0.94266653]),
  array([ 0.15608799, -0.31515813,  1.52390297]),
  array([-0.64049246, -0.55042197, -0.03932225]),
  array([-0.7351579 , -0.42261309,  0.62528872]),
  array([-0.47112498, -0.43855884,  1.79611997]),
  array([-1.14771865, -0.91790413,  0.29107333]),
  array([-0.86978653, -0.99849349, -1.42922211]),
  array([-0.50010667, -0.46019107,  1.25023328]),
  array([-0.86859682, -1.01124499, -0.56536951]),
  array([-1.15039383, -1.47472448,  0.63972234]),
  array([-0.62764249, -0.67006046,  0.8280986 ]),
  array([0.04191571, 0.40843992, 1.5692791 ]),
  array([-0.29269049, -0.42675805,  0.28849878]),
  array([-1.08948177, -0.93330622, -0.25143828]),
  array([-0.09793955, -1.626645  ,  1.26453917]),
  array([-0.92209821, -1.25493864, -0.59451982]),
  array([-0.80578418, -0.08549119,  2.22183844]),
  array([-0.81021861,  0.16890828, -0.04039344]),
  array([-1.08856527,  0.5903738 ,  1.20987478]),
  array([-0.82613672, -1.15382122, -0.33690825]),
  array([-1.05386056, -1.05931393, -0.33675768]),
  array([-1.2849929 , -0.05612537,  2.08396479]),
  array([-0.86784329,  0.11844515,  0.91861897]),
  array([-0.23719832, -0.09847393, -0.52617732]),
  array([-0.96048834, -1.07970956, -0.51429085]),
  array([-0.61935258, -0.35837702, -0.25359892]),
  array([-0.28553194, -0.36843062, -0.07881956]),
  array([-0.56543456, -1.08642657,  0.77105181]),
  array([-0.2487892 , -1.01786106, -0.77142037]),
  array([-0.98011984, -1.7025568 , -0.60680165]),
  array([-0.46668071, -0.64440045, -0.00918324]),
  array([-0.54867841, -0.48502869, -0.70859229]),
  array([-1.16344414, -0.31345349,  0.84609255]),
  array([-0.91458622, -1.2090945 ,  0.28025755]),
  array([-0.64714106,  0.11190112, -0.23073433]),
  array([-1.1776213 , -0.60579174,  1.04367464]),
  array([-0.87446896, -0.57874938,  0.00935885]),
  array([-1.12118677, -0.07053793,  1.37671067]),
  array([-0.85823234, -1.3775795 ,  1.5352938 ]),
  array([-1.19314306, -1.09781962,  0.35859546]),
  array([ 0.13812179, -0.48909517, -0.11363494]),
  array([-0.8007479 , -1.24067163, -0.20221921]),
  array([-0.34226783, -0.01238195, -0.01412357]),
  array([-0.82814397, -0.27689924,  0.18259718]),
  array([-0.6743343 , -0.70581195,  0.07810142]),
  array([-1.01745455, -0.46320764, -1.18954939]),
  array([-0.63366615, -0.1721056 ,  0.24042972]),
  array([-1.25226808,  0.06400203,  1.14917502]),
  array([-1.02202783, -0.1977732 ,  0.11687091]),
  array([-0.27791368,  0.58231359,  1.38203871]),
  array([-0.96062954,  0.10627662, -0.50574387]),
  array([-0.45865183, -1.62493563,  1.07220677]),
  array([-0.58610094, -1.41373386,  1.23663335]),
  array([-1.26055594, -0.13432897,  1.62834489]),
  array([-0.77423153, -0.33001603, -0.69597824]),
  array([ 0.39558933, -1.38596152,  1.10077481]),
  array([-0.89665058, -1.40259402,  0.09844436]),
  array([-0.95982486, -1.16241483,  0.19429903]),
  array([-1.154444  , -1.7559617 ,  0.18647637]),
  array([-1.10634831, -0.56224637,  0.43928845]),
  array([-0.8039507 , -0.91514347,  0.28799452]),
  array([-0.82006964,  0.28880077,  2.02438599]),
  array([-0.66640213, -0.65394817,  0.09979005]),
  array([-0.30261478, -0.50414634, -0.8543424 ]),
  array([-1.01643405, -0.029601  ,  0.11032734]),
  array([-0.5261763 ,  0.13722595,  0.18119013]),
  array([ 0.08974389, -0.76837379, -0.13167758]),
  array([ 0.19455304, -0.68233199, -0.23910487]),
  array([-1.00877078, -0.14334698, -0.01085648]),
  array([ 0.20620874, -0.13274476,  0.18369814]),
  array([-0.98784694, -1.38503516,  0.23304211]),
  array([-0.67414011, -0.36935959, -1.35348521]),
  array([-0.53114684, -0.25363288, -0.80202607]),
  array([-0.81230448, -0.00219722, -0.08508945]),
  array([-0.9106356 , -1.53813409, -1.59377009]),
  array([-0.45118518, -0.34374925, -0.5621723 ]),
  array([-0.92257946, -0.43644967, -1.04676935]),
  array([-0.56551525, -1.40582821, -0.13649282]),
  array([-1.09532057, -1.77159826, -0.36429396]),
  array([-0.53051024, -1.34646945, -0.31026534]),
  array([-0.86612947, -0.43506772, -0.18209348]),
  array([-0.82408246, -0.4132211 ,  0.91914886]),
  array([-0.39597734, -0.9031579 , -0.70768388]),
  array([-1.01794654, -0.12762364, -0.47575951]),
  array([-1.08055721, -0.85327754,  0.91890178]),
  array([-0.01877395, -0.30979239, -0.33714735]),
  array([-0.72252038, -0.79279775, -1.45600407]),
  array([-0.89719191, -1.16497021,  1.84018714]),
  array([-0.47928373, -0.3747609 , -0.14710744]),
  array([-0.84912867, -1.3052623 , -1.95265405]),
  array([-0.13723442, -0.35515915, -0.14094896]),
  array([-0.78227311, -0.14305056,  0.66047391]),
  array([-1.25223289, -1.19865206,  1.80250464]),
  array([-0.66212069, -0.71182431, -0.039408  ]),
  array([-0.67448567, -0.61985078, -0.84224804]),
  array([-1.24390338, -1.69913244,  1.29672992]),
  array([ 0.47198367, -0.97351208, -1.45244737]),
  array([-0.16821307, -0.769784  , -0.03356644]),
  array([-1.10701471, -0.90462918, -0.50653732]),
  array([-0.97908295, -1.22565964,  0.14388594]),
  array([-1.11426528, -0.06898027,  0.69839305]),
  array([-0.79295067, -0.40680757,  1.40668621]),
  array([-0.70097701, -0.19609858, -0.4800398 ]),
  array([-0.90011527,  0.44309372,  0.65598152]),
  array([-0.78740327, -0.70750666,  0.13416394]),
  array([-0.3535521 , -1.30675085,  1.20501065]),
  array([-1.23324989,  0.2441459 ,  2.07987254]),
  array([ 0.36206698, -0.51958489, -1.04147578]),
  array([-0.39020866, -0.32532045, -0.01618676]),
  array([-0.90577839, -0.53880741,  0.67653276]),
  array([-0.62348342, -1.20404798,  0.11833679]),
  array([-0.61428511, -0.04554292,  0.37290467]),
  array([-0.04432281, -0.29028616,  0.03332165]),
  array([-0.53323544, -0.99348943, -1.60514183]),
  array([-0.83701239, -1.06500674, -0.29428881]),
  array([-0.27238777, -0.04049269,  0.99202258]),
  array([-0.87835656, -1.32311988, -0.44282109]),
  array([-0.43475256, -0.53311123,  0.67338529]),
  array([-1.26726528,  0.43356722,  0.60952125]),
  array([-0.93723001, -1.61765805,  1.4267888 ]),
  array([ 0.02615347, -0.57949281, -0.78671404]),
  array([-0.72889593, -0.67718878,  0.46598679]),
  array([-0.09027552,  0.39245299,  1.10299893]),
  array([-0.44252786, -0.06010628, -0.05742089]),
  array([-1.09612719, -0.47557049, -0.73355178]),
  array([-0.16514143, -1.4411948 , -0.19556611]),
  array([-1.11528688, -1.73314805,  0.09858976]),
  array([-0.28842167, -0.41076094, -0.69613642]),
  array([-1.14095975, -0.18140511,  0.17506617]),
  array([-1.30983692, -1.71486444,  1.63405504]),
  array([-0.80238884, -0.02766932,  1.83948637]),
  array([-0.68241768, -0.75286317, -0.78151484]),
  array([-0.94246205, -0.80319457,  0.23823662]),
  array([-1.06366883, -1.77996929, -0.65914903]),
  array([-0.71056105, -1.12376628, -0.91339149]),
  array([-0.7912641 , -0.03811243,  1.18184492]),
  array([-1.02249279, -0.84364897,  0.60314035]),
  array([-1.15858197, -1.75486731,  0.22502409]),
  array([-0.93482091, -0.28312952,  0.13592101]),
  array([-0.8570618 , -1.00971056, -0.68693054]),
  array([ 0.14228855, -1.03027507, -0.0558529 ]),
  array([-1.19861672, -0.41529512,  0.4980854 ]),
  array([-0.66812151, -0.32167588, -0.29858058]),
  array([-1.08225668, -1.27617233,  0.30272687]),
  array([-0.90758642, -0.17594625,  0.19787793]),
  array([-0.09845241, -0.21230111, -0.17994519]),
  array([-0.84974077, -0.673038  ,  0.16437649]),
  array([ 0.01584711, -0.82073103, -0.25118739]),
  array([-1.19356231,  0.04587779,  1.81959971]),
  array([-0.83660317, -0.60799623,  0.18533488]),
  array([0.12126256, 0.35994716, 1.89702757]),
  array([0.03341639, 0.43160216, 1.62459857]),
  array([-0.5729984 , -0.57478662,  0.05729108]),
  array([-0.76241869, -0.31320076, -0.00673775]),
  array([-0.41212033, -0.08411316, -0.1179215 ]),
  array([-1.28944027,  0.04066342,  1.21857081]),
  array([-1.12513064, -1.51457557,  0.12749492]),
  array([-0.947969  , -0.9294934 ,  1.23875176]),
  array([-0.93621132, -0.51860796, -0.01597663]),
  array([-1.33120194, -1.70921397,  1.83308314]),
  array([-0.97423834, -0.27357854, -0.54118629]),
  array([-0.57213801,  0.74650722,  1.93197646]),
  array([-0.5072968 , -0.27900963,  0.54044082]),
  array([-0.72773671, -1.63257737,  1.26647592]),
  array([-1.02744633, -0.84960603,  1.92077024]),
  array([-0.81716595, -0.7908751 ,  1.18874257]),
  array([-0.62690028, -0.88818158,  0.22699908]),
  array([-0.14449453, -0.09048569, -0.03754567]),
  array([-0.57264795, -0.67929744,  0.33756936]),
  array([-0.48461092, -1.41851085,  0.63253213]),
  array([-1.0459543 , -1.03723814, -0.18190003]),
  array([-0.74998821, -0.97066739, -0.33112279]),
  array([-0.71701188, -0.22480456,  0.22388527]),
  array([-1.05081186, -1.63147488,  1.42632759]),
  array([-0.56892832,  0.04075087, -0.30198038]),
  array([-0.89105455, -1.40961219,  0.13085439]),
  array([-0.94513031, -0.26501033, -0.13218689]),
  array([-0.63130062, -0.84677304, -0.56841521]),
  array([-1.07686132, -0.27863207,  0.95410062]),
  array([ 0.20830529, -0.067617  ,  1.31735451]),
  array([-0.78066474, -0.31548559, -0.7092025 ]),
  array([ 0.26474187, -0.37705211,  0.35151395]),
  array([ 0.12753331, -0.72917022, -0.13630177]),
  array([-1.00185976, -1.41677613,  0.48814532]),
  array([-0.26125683, -0.75034135, -0.31017398]),
  array([-0.99612189,  0.01003   ,  0.04170856]),
  array([-0.6981589 , -0.24260859,  0.89624102]),
  array([-0.1592147 , -1.33461439, -0.1536839 ]),
  array([-0.47248526, -0.90702413,  0.87980557]),
  array([-1.1283427 ,  0.16067005,  1.0576789 ]),
  array([-0.17911671, -0.43030941,  0.03824726]),
  array([-0.81101294, -0.63314782,  0.74671625]),
  array([-1.10423971, -1.26779549,  0.54497186]),
  array([-1.08852344,  0.53926678,  0.88933607]),
  array([-0.43221469, -1.02050453, -2.21746684]),
  array([-0.28226194, -0.92061422,  1.43962516]),
  array([-0.93097602,  0.2382842 ,  0.12302541]),
  array([-0.41884829,  0.01368132,  0.57264549]),
  array([-0.79123601, -1.21662446, -1.36100563]),
  array([-0.83165747, -0.96832418, -2.37634149]),
  array([-0.2087513 ,  0.14923212,  0.96619539]),
  array([ 0.0668723 , -0.77727218, -0.08710325]),
  array([-0.5342697 , -1.32440016, -0.0179428 ]),
  array([-0.83133854, -1.53012119, -0.124446  ]),
  array([ 0.10703139, -1.19174507,  0.92146932]),
  array([-1.05368076, -1.17178408, -0.00775399]),
  array([ 0.08626361, -0.57413976, -0.03574538]),
  array([-1.12027228, -1.7649992 , -0.13185378]),
  array([-0.91944408, -0.03107554,  0.29345937]),
  array([-1.14182057, -0.89542397, -0.18229982]),
  array([-1.02936899, -0.77768447,  0.12069077]),
  array([-0.58065155, -1.48616728, -2.84527701]),
  array([-0.82551899, -0.96111724,  0.17677427]),
  array([-1.01171512,  0.40048753,  0.48569915]),
  array([-0.78547539, -1.23953908,  0.00914218]),
  array([-1.10002592, -0.03844365,  1.63343766]),
  array([ 0.30724123, -0.48658489, -0.23113847]),
  array([-0.9953844 , -1.1095931 , -0.54751577]),
  array([-0.20657191, -1.35840224,  0.66224499]),
  array([-0.87414979, -0.40196246,  0.58543055]),
  array([-0.55589609, -1.74780616, -0.51576343]),
  array([-0.78460869, -0.6429994 , -0.22687316]),
  array([-0.61599272,  0.39406747,  1.48204635]),
  array([-0.6858823 , -1.73473667, -0.07953243]),
  array([-0.7680403 , -0.02137298,  0.03056667]),
  array([-0.37781993, -0.23954455, -0.09857555]),
  array([-1.19766836,  0.23236129,  0.30086492]),
  array([-1.16825869, -0.61712309,  0.45899493]),
  array([-0.98553454, -0.48108234,  0.19911423]),
  array([-0.55397188, -0.44452384, -0.11378661]),
  array([ 0.31311571, -1.29147375, -0.61546564]),
  array([-0.74242386,  0.22321267,  0.11328081]),
  array([-1.00397458, -1.66307821, -0.10769491]),
  array([-0.55530632,  0.10047664,  0.03554512]),
  array([ 0.0153936 , -0.51816342,  0.6632836 ]),
  array([-1.03615731, -0.31714685,  1.93441629]),
  array([-0.70712383, -0.89410515, -0.16314339]),
  array([-0.73896007, -0.69862563, -0.1557353 ]),
  array([-0.57775092,  0.29488498,  1.6023999 ]),
  array([-1.1271759 ,  0.54841682,  1.1513222 ]),
  array([-0.66291255, -0.42333868,  0.2606138 ]),
  array([-0.80085657, -1.61524459,  0.42977849]),
  array([-0.82472112, -0.36194687, -0.07076067]),
  array([-0.99563685,  0.02557514, -0.18585906]),
  array([-1.07545415, -0.8196282 , -0.10164317]),
  array([-0.90077904, -1.21222248, -1.43212568]),
  array([-0.91291157, -1.21459223,  0.00570135]),
  array([-1.06366501, -1.4334671 ,  0.7870555 ]),
  array([-0.63749382, -0.19188322,  0.33314995]),
  array([-0.92397846, -0.11596309,  0.18786746]),
  array([ 0.11739618, -1.35438437, -0.1339605 ]),
  array([-0.64229192, -1.13221568, -0.53106709]),
  array([-1.18775427, -0.88327574,  0.24560034]),
  array([-1.37242279, -0.80126622,  2.24278437]),
  array([-0.08191015, -0.34770956, -0.07123717]),
  array([-0.91918203, -0.28912328,  0.02242992]),
  array([-0.03057864, -0.58496379, -0.26097779]),
  array([ 0.2227294 , -0.76980427, -0.32354354]),
  array([-0.55497492, -0.41238276, -0.51373043]),
  array([-0.99254608, -0.43629151, -0.15029132]),
  array([-1.02692144, -0.15168268, -0.62075871]),
  array([-1.13379714, -1.72825258,  0.27102397]),
  array([-1.01997833,  0.01895616,  0.39393239]),
  array([ 0.171061  , -1.39948557, -1.21758144]),
  array([-0.94571255, -0.45647081,  1.78321087]),
  array([-0.18810738,  0.00135199,  1.00446795]),
  array([-0.98281977,  0.04012506,  0.26950638]),
  array([-0.82178741,  0.18992571,  0.22201282]),
  array([-1.09350105,  0.2900097 ,  1.39729923]),
  array([-0.86862805,  0.43422056,  1.32086951]),
  array([-0.8645661 , -0.44488306, -0.53717922]),
  array([ 0.28775475, -0.54996766,  0.23595874]),
  array([-0.84130285, -0.97221521, -0.09061499]),
  array([-0.50454658,  0.50150936,  1.01026787]),
  array([ 0.0310844 , -0.66704989, -0.0522492 ]),
  array([ 0.20783645, -1.55310223, -1.20119135]),
  array([-0.32283291, -1.59978283, -0.24128458]),
  array([-0.83626244, -0.3192719 , -1.53343586]),
  array([-0.95881224, -1.19091765, -0.5370986 ]),
  array([-1.0349041 , -1.73137889,  1.2033869 ]),
  array([ 0.14399645, -1.5350445 ,  0.03367231]),
  array([-0.31257281, -1.35893089,  1.77652907]),
  array([-0.76947019, -1.45259994,  0.82256308]),
  array([-0.97843871,  0.05564842,  1.74558988]),
  array([-0.58284252, -0.75426758,  1.63044993]),
  array([-0.43515166, -0.87285028, -0.38331265]),
  array([-1.10182297, -1.70353924,  0.25005083]),
  array([-0.96951659, -1.6757727 ,  1.86298514]),
  array([-1.04637517, -0.88749706, -0.79454565]),
  array([-0.84351434, -0.04318332, -0.47835535]),
  array([-0.91917009, -0.97986399,  0.00250352]),
  array([-0.79853422, -1.18843129, -1.41993519]),
  array([-1.28470075, -0.70875056,  2.13985863]),
  array([ 0.22406313, -1.57699554, -0.24623435]),
  array([-0.64074679, -0.13758746, -0.53443014]),
  array([-0.90651435, -0.80534297,  0.2698813 ]),
  array([-1.05313861, -0.4772067 , -0.07816334]),
  array([-0.74776744,  0.46064716,  1.04948589]),
  array([-0.13891408, -0.60790765, -0.5877603 ]),
  array([ 0.54096201, -1.47437792, -0.19830856]),
  array([-0.94696096, -0.64062411,  1.71202214]),
  array([-0.61777745, -1.7394469 ,  0.02669141]),
  array([-0.992311  , -0.37291833, -0.0777912 ]),
  array([-0.73478517, -1.22310729,  1.24316087]),
  array([-0.92213105, -1.27102797,  0.62809901]),
  array([-1.06366883, -1.77996929, -0.65914903]),
  array([-0.85418225, -0.56123533,  1.61896646]),
  array([-0.24529221, -0.00481364, -0.11050667]),
  array([ 0.28773618, -0.36335364,  1.53887492]),
  array([-1.2858136 , -0.60820295,  1.37317222]),
  array([-0.21342558, -0.91924363, -0.09332483]),
  array([ 0.30846684, -1.21411743, -2.81798573]),
  array([-1.33343627, -0.84474676,  1.60271591]),
  array([-0.91421203, -0.74563234,  0.27380101]),
  array([-0.08142462, -0.53692145, -0.17049499]),
  array([ 0.37562114, -1.60440592, -0.18128719]),
  array([-0.60549942, -1.66198283,  0.08421895]),
  array([-0.8194566 ,  0.34677889,  1.02713205]),
  array([-0.8558263 , -1.71998294, -1.4255027 ]),
  array([-0.45891963,  0.0648461 , -0.12940414]),
  array([ 0.2708774 , -0.83150135,  1.07328417]),
  array([-0.63659545, -1.34464961, -1.5709967 ]),
  array([0.00113267, 0.2799844 , 0.84129088]),
  array([-1.25862944, -1.04733058,  1.24552608]),
  array([-1.11817475, -1.26727649,  0.27678686]),
  array([-1.16305723, -1.75368373,  0.26671383]),
  array([-0.76818523, -0.78210231, -0.57172254]),
  array([-0.89432472, -0.48785599,  0.17227504]),
  array([-1.00180382, -0.83723464,  0.1470032 ]),
  array([-0.83219501, -1.16011103, -2.72696766]),
  array([ 0.56910045, -1.73099697, -0.75565583]),
  array([-0.74800514,  0.00700318, -0.39971312]),
  array([ 0.14401207, -0.99238617, -1.34983465]),
  array([-1.07388233, -1.27899065, -0.13582372]),
  array([-1.24239292,  0.21102008,  0.4406156 ]),
  array([-0.67828684,  0.05441717,  0.09022948]),
  array([ 0.31234028, -0.65980261, -0.34851939]),
  array([-0.73567109, -0.49859348, -0.14656116]),
  array([-0.51495149, -0.21148017, -0.04046812]),
  array([-0.88882674,  0.33531704,  0.47846912]),
  array([-0.8799247 ,  0.30759454,  0.15085815]),
  array([-0.08241756, -0.1856605 , -0.14890506]),
  array([-0.14242454, -0.62877023, -0.31319209]),
  array([-0.00655009, -0.28920286,  0.26772855]),
  array([-0.90667452, -0.03545758,  1.27088417]),
  array([-0.48181963, -1.44187852, -0.01421788]),
  array([-0.91914357, -0.45783492, -0.47470945]),
  array([-1.2148937 , -1.73997437,  0.74960181]),
  array([-0.95737388, -0.45538107, -0.68784128]),
  array([0.02541956, 0.15158543, 1.47639993]),
  array([-1.17152517, -1.71827453,  0.6224833 ]),
  array([-0.46083007, -1.00818494, -0.03022065]),
  array([-0.68912933, -1.17080626, -0.38774744]),
  array([-0.80102534, -0.00430612, -0.08131242]),
  array([0.22458886, 0.09319805, 0.6916644 ]),
  array([-0.95108944, -1.39475653, -0.10937619]),
  array([-0.70280282, -0.93847826, -0.9768134 ]),
  array([-0.63653377, -1.69289517, -2.17736469]),
  array([-0.71596436, -0.90752058, -2.51925872]),
  array([-0.57109229, -0.51565228, -0.08743484]),
  array([-0.16113557, -0.24841629,  1.03450938]),
  array([-1.01249207, -0.49768974, -1.51266333]),
  array([-1.25911498,  0.11983004,  1.50321559]),
  array([-0.39299163, -0.02772095, -0.28139965]),
  array([-0.87358686,  0.62839066,  1.29885884]),
  array([-0.90929   , -1.03388581,  1.5372609 ]),
  array([-0.63258986, -0.4605111 ,  0.44510782]),
  array([-0.98242039, -1.5523186 , -1.201937  ]),
  array([-0.89629763, -0.1757964 , -0.08287391]),
  array([0.18122827, 0.09740223, 0.64196858]),
  array([-0.5216074 , -0.15391307, -0.82694355]),
  array([ 0.35098992, -0.65151499, -1.30382957]),
  array([-0.979497  , -0.27218777, -0.49219866]),
  array([ 0.11082574, -0.6879081 , -0.0825978 ]),
  array([-0.58135866, -1.22585576,  1.64877198]),
  array([-0.89644288, -0.64334912, -0.54029192]),
  array([-0.15061002, -0.08161151, -0.39360688]),
  array([-0.33382071, -1.05605329, -0.05378617]),
  array([-0.83027765, -0.35470395,  0.0532834 ]),
  array([-1.05612319, -1.74879526, -0.45255571]),
  array([-0.12062839, -0.36796551,  1.81113654]),
  array([-0.82625901, -0.22957941, -0.82991895]),
  array([-0.95374451, -0.22951337,  0.35590401]),
  array([-0.26250731, -1.18612486,  1.3946259 ]),
  array([-0.76204656,  0.3717862 ,  0.76910295]),
  array([-0.35201103, -0.85816487,  0.04286587]),
  array([-0.95903848, -0.46105445,  0.15022595]),
  array([ 0.07601851, -1.18952213,  0.06468029]),
  array([-0.47309023, -0.22426256, -0.48286462]),
  array([-1.04962179, -1.17285757, -0.04556577]),
  array([-0.95291248,  0.2030267 , -0.00602566]),
  array([ 0.2269007 , -0.90204657, -1.93099516]),
  array([-0.80442558, -0.39355886,  0.34713419]),
  array([-1.10307113, -1.76954844, -0.29209282]),
  array([-0.96313099, -1.09622311, -0.02062318]),
  array([-0.58888994, -1.70717567,  0.06018269]),
  array([-0.39337086, -0.15325263, -0.72108556]),
  array([-0.82675776,  0.54178476,  1.36564852]),
  array([-0.81642942, -0.05497138, -0.08315129]),
  array([-1.06782375, -1.71253112, -0.06667238]),
  array([-0.89629405, -0.99255732,  0.1433671 ]),
  array([-0.65669819, -0.45583366, -1.85860615]),
  array([-1.26603013, -0.7748318 ,  2.22213904]),
  array([ 0.5852919 , -1.58635526, -0.50523589]),
  array([ 0.2683355 , -0.92916039,  0.87797077]),
  array([-0.86137389,  0.09672577,  0.18181152]),
  array([-0.60001005, -0.12256431, -0.69333769]),
  array([-0.15195394, -0.10771637,  0.92963823]),
  array([-1.04797556, -1.69852215,  0.84350413]),
  array([-1.00138684, -1.74680343,  0.17620221]),
  array([-0.86785166, -1.35043138, -0.33101523]),
  array([-1.02433963, -0.95430778,  0.66825717]),
  array([-0.00148547, -1.18686583, -0.11960456]),
  array([-0.20343646, -0.51355284, -0.13931619]),
  array([-0.37406886, -0.02228348,  1.63079877]),
  array([-1.19788082, -1.74447382,  0.59111656]),
  array([-0.14922263, -0.77958669,  0.00923189]),
  array([-1.30937789, -1.71498584,  1.62977894]),
  array([-0.01268401, -1.05184145,  0.48012725]),
  array([-0.96427463, -0.48772054,  0.24283517]),
  array([-1.06366883, -1.77996929, -0.65914903]),
  array([-0.79664123, -0.04282032, -0.13669503]),
  array([-0.86809496, -1.35036703, -0.32874872]),
  array([-0.2990024 , -1.05285302,  0.07307531]),
  array([-0.26801863, -0.74124758, -0.2179006 ]),
  array([-1.36969446,  0.46065699,  1.56371083]),
  array([ 0.24938789, -0.32452218, -0.62399353]),
  array([-1.17054701, -1.25342544,  0.76466594]),
  array([-0.61442855,  0.10434563, -0.22279739]),
  array([-0.76698897, -0.18592588, -0.30533702]),
  array([-0.93481693,  0.04698758, -0.46679331]),
  array([-1.13432371,  0.00876886,  0.46888105]),
  array([-1.26955049, -1.72551912,  1.25876275]),
  array([-1.19962942, -0.5792141 ,  1.58436664]),
  array([-0.87289031, -0.76612138,  1.89722337]),
  array([-1.17334881, -0.45514745, -0.01418593]),
  array([-0.56144378, -0.39176559,  1.04057295]),
  array([-0.75217571,  0.09585275,  0.08042751]),
  array([-0.97242563, -1.47317648, -0.40210501]),
  array([-0.99754014, -1.44881087, -0.23043092]),
  array([-1.21762272, -0.44343819,  0.3982523 ]),
  array([-0.82924893, -1.09865202,  0.08398389]),
  array([-0.8834913 , -1.72155233, -0.79200349]),
  array([0.04536219, 0.43441212, 1.65716367]),
  array([-0.89897682, -0.23422957,  0.21514734]),
  array([-0.66887518, -0.29272592,  0.81233733]),
  array([-0.61109869, -1.36428223, -0.19083366]),
  array([-0.42228775, -0.28423623, -0.75337486]),
  array([-1.28407382, -1.72167809,  1.39405627]),
  array([-0.03715734, -0.65516285,  1.54893629]),
  array([-0.44436396, -1.44387465, -0.9302193 ]),
  array([-1.08265436, -1.52580941, -0.26819723]),
  array([-0.54159776, -0.28908115, -1.18287147]),
  array([-1.25609746, -0.60061592,  1.43551665]),
  array([-0.80022791, -0.38355739, -0.22220922]),
  array([-0.94395996, -1.31335163, -1.34612896]),
  array([-0.21100059, -0.2782964 ,  1.10685132]),
  array([-1.15947924, -1.72146036,  0.51026816]),
  array([-0.48011785, -0.40849369, -0.09152952]),
  array([-0.70236824,  0.65835529,  1.71612189]),
  array([-0.97159206, -0.89709756,  1.08207656]),
  array([-0.63537398, -0.85673233,  0.71246889]),
  array([-0.48485513, -0.33972784,  0.51083102]),
  array([-0.41641129, -1.28149022,  0.08734204]),
  array([ 0.15140021, -0.61754092,  0.08250483]),
  array([-1.14250181,  0.25094087,  0.06384075]),
  array([-1.14265318, -0.31593376,  0.28108128]),
  array([-0.47161818, -1.1532939 ,  1.8442431 ]),
  array([-0.20778667, -0.07863665,  1.56640098]),
  array([-1.21983997,  0.01366212,  1.32779927]),
  array([-1.14112247,  0.31895609, -0.00560253]),
  array([-1.01179935, -0.41129556,  0.93185129]),
  array([-0.90125278, -0.51178858, -0.8698972 ]),
  array([-0.0683246 , -0.50826182, -0.60252189]),
  array([-1.00229065,  0.11918957, -0.19725593]),
  array([-1.08757848, -1.20002621,  0.59391137]),
  array([-0.65450779, -0.46387715, -0.48121894]),
  array([-0.31867998, -0.62159222, -0.43019725]),
  array([-0.36869681, -0.46200913, -0.93251224]),
  array([-0.83328189, -0.71997248, -0.57170279]),
  array([-0.77423148, -0.51721884,  0.14779006]),
  array([-0.32406907, -0.93520972,  0.63670112]),
  array([-0.61981637, -0.01900384, -0.23077044]),
  array([-0.60416715, -0.49488501,  1.48874489]),
  array([ 0.00398122, -0.44332093,  1.267522  ]),
  array([-0.85760067, -0.76883275,  0.70976769]),
  array([ 0.15776525, -1.19777219, -0.40495259]),
  array([-0.74905588,  0.14794951,  1.6978733 ]),
  array([-0.75619611,  0.28917125,  0.32257357]),
  array([-0.87946637, -0.68171244, -0.89837999]),
  array([-1.18284123, -1.17079426,  0.91856808]),
  array([ 0.19666961, -0.4204197 , -1.01228341]),
  array([-1.25143018, -0.9980592 ,  1.55209867]),
  array([-0.52012153, -1.25908073, -0.84975933]),
  array([-0.94182583, -1.29774953, -0.24214359]),
  array([-0.88981536, -1.71261563,  0.8597242 ]),
  array([-0.75243781, -0.56649462, -0.18832658]),
  array([-0.68180253, -0.43721183, -0.02633732]),
  array([-0.52181501, -0.91454036, -0.65626411]),
  array([-0.5283169 , -1.43725653,  1.43203118]),
  array([-0.8151633 , -1.5260259 , -1.00877647]),
  array([-0.96114218, -0.82786529, -1.23758117]),
  array([-0.99008891,  0.01141381,  1.62139267]),
  array([-0.52802935, -1.65970928, -0.01667138]),
  array([-1.04100376, -0.28909054, -0.19611171]),
  array([-0.82267701, -0.38803678, -0.73195381]),
  array([-1.04043917, -0.92335348,  0.39456986]),
  array([-0.44911233, -1.48550505,  0.16615519]),
  array([-0.92662797,  0.40532411,  0.4348878 ]),
  array([-0.97671237, -0.58968198, -0.53162412]),
  array([-0.86701681, -1.02705403, -0.67917581]),
  array([-0.68389033, -0.32681078, -0.80161361]),
  array([0.02893261, 0.02995029, 0.03158551]),
  array([0.09576678, 0.26269405, 1.4744666 ]),
  array([-0.73331757, -0.93187976, -0.41562937]),
  array([-0.96223222, -0.24763887, -0.37990907]),
  array([-1.03066683, -1.53955871, -0.75249238]),
  array([-0.96316562, -0.48804907,  0.16408055]),
  array([-1.32397666,  0.01662772,  1.26341267]),
  array([-0.68008344, -0.30194412,  1.43122338]),
  array([-0.6826732 , -0.35048183,  0.05290005]),
  array([-0.80999229, -0.54109393, -1.90426455]),
  array([-0.00384045,  0.00849962,  0.5861312 ]),
  array([-1.27454998,  0.80280606,  1.59142953]),
  array([-0.50494493, -0.83794389, -0.56912027]),
  array([-1.02350403, -1.7242525 , -0.47953733]),
  array([-0.68987709, -0.74015799, -0.11078744]),
  array([-0.191064  , -0.30113862, -0.38485965]),
  array([-0.00749478, -1.10776924, -0.49652626]),
  array([-1.22269087, -1.48877351,  1.03632725]),
  array([-0.54437662, -1.15899397,  0.02504929]),
  array([-1.12679532, -0.43428995, -0.1709742 ]),
  array([-0.95349366, -0.76914371,  0.91856405]),
  array([-0.86199383, -0.39414239, -0.03680129]),
  array([-0.04390779, -0.10178263,  0.37563298]),
  array([-0.35025117, -0.11415974, -0.59125316]),
  array([-0.85887997,  0.30891906,  0.31700276]),
  array([-0.90705116, -1.82139046, -2.11813702]),
  array([-1.01146519, -1.21360874, -0.11965031]),
  array([-1.0990006 , -0.31755463,  0.04282916]),
  array([-0.16041709, -1.56420295, -3.1814072 ]),
  array([ 0.16025261, -1.11683112, -1.07440852]),
  array([-0.88358844, -0.82787236,  0.03085107]),
  array([-1.0242921 , -0.98373572,  2.14475604]),
  array([-0.40091221, -0.42286536,  1.10106163]),
  array([-0.81730821, -0.4802006 , -0.01782521]),
  array([-0.85005198, -1.42045626, -0.25110924]),
  array([ 0.14449867, -0.29917153, -0.13238852]),
  array([-1.01619445, -0.40729372, -0.24289205]),
  array([ 0.20616627, -1.35106808, -0.17330348]),
  array([-0.75316693, -0.21728546, -0.31449168]),
  array([ 0.17204395, -1.72228648, -0.00992649]),
  array([-1.33107543, -0.23029947,  2.10986176]),
  array([-1.17891123, -0.94041841,  1.1583306 ]),
  array([-0.97028472, -1.69329304,  1.93242347]),
  array([-0.94352007, -1.23382698, -0.0557942 ]),
  array([-0.70831655, -0.60076949, -0.10348007]),
  array([ 0.3411963 , -0.37625754, -0.45290393]),
  array([-0.93032728, -0.55185837, -0.05649061]),
  array([-0.73537654, -1.42817503,  0.09527788]),
  array([-1.11479688, -0.6312623 , -0.03903407]),
  array([-0.91182011,  0.31738091,  0.2426941 ]),
  array([-0.91301999, -1.06554998, -0.26640124]),
  array([-0.05558246, -1.37335359, -1.0449296 ]),
  array([-0.69484448, -0.08853521, -0.26039099]),
  array([-0.1347646 , -0.97380128, -0.11765638]),
  array([-3.42855398e-02,  3.70799916e-04,  4.65810909e-01]),
  array([-0.12862094, -0.59746543, -0.4837464 ]),
  array([-0.82058659, -1.26438449, -2.10744324]),
  array([-0.64486698, -0.6592403 , -0.13314202]),
  array([-1.12493552,  0.70535106,  1.31863958]),
  array([-0.41698104, -0.60861572,  0.07489865]),
  array([-1.27838787, -1.72318187,  1.34108824]),
  array([-0.27803951, -0.49576132, -0.19421278]),
  array([ 0.30595745, -0.98673531, -0.22592407]),
  array([-0.90763031, -1.38929915, -2.23833254]),
  array([-0.08221995, -0.68435713, -1.11394601]),
  array([-0.75391021, -0.79330503, -0.76068964]),
  array([-0.87972978, -0.58127534,  0.43343806]),
  ...],
 [array([1.31262106, 0.15646992, 0.04335054]),
  array([0.90423265, 0.66147204, 1.51477739]),
  array([ 1.38943714, -0.46667871, -0.18455536]),
  array([0.65392575, 0.00596636, 0.16657004]),
  array([1.95444096, 0.17401042, 0.73515143]),
  array([ 1.25644789,  0.27090637, -0.22639553]),
  array([ 0.63738417, -0.82159773,  0.31265212]),
  array([ 1.2709252 , -0.309183  , -0.26194284]),
  array([1.77500467, 0.66091516, 0.69635924]),
  array([ 0.8992651 , -0.3772873 , -0.81062758]),
  array([ 0.4341125 , -0.62400555,  1.610751  ]),
  array([ 1.87110137, -1.37028659,  0.72435594]),
  array([ 1.88514513, -0.94823303,  1.32005024]),
  array([1.06456642, 0.13451338, 1.19450455]),
  array([ 1.69155493, -0.57893651,  1.34163557]),
  array([ 1.26227883, -0.38803931,  1.45868449]),
  array([ 2.78414194, -0.68040554, -0.63583978]),
  array([ 0.33813852, -0.05590023,  1.61311387]),
  array([2.15598652, 0.26714268, 0.911236  ]),
  array([ 0.85922185, -0.7156157 , -0.47444424]),
  array([ 1.26113156, -1.2671849 , -1.41189411]),
  array([ 2.81923915, -0.88184764,  0.05384373]),
  array([ 2.22057062,  0.29996406, -2.47882769]),
  array([ 1.21828445,  0.16566239, -0.62890639]),
  array([ 2.07184895, -0.0585719 , -0.37771045]),
  array([ 2.04300401, -0.5362357 , -0.58879309]),
  array([ 1.35284339, -0.50909743, -1.18503912]),
  array([ 1.12463868,  0.13815496, -0.96524423]),
  array([1.60559146, 1.51187159, 0.75317287]),
  array([ 0.44691822, -0.08065661, -0.31495127]),
  array([ 3.03624952, -1.32784463,  1.40608587]),
  array([0.24730024, 0.45117905, 1.35342933]),
  array([ 0.73421634, -0.42960841,  1.48293992]),
  array([ 0.52844367,  0.13557442, -0.03034484]),
  array([1.49729239, 1.05338133, 0.18997306]),
  array([ 0.851603  , -0.34188661,  0.81742183]),
  array([ 1.31653973,  0.5160052 , -1.0669596 ]),
  array([1.52885287, 0.23551746, 1.32047842]),
  array([ 1.34134972, -0.46470047, -0.30944175]),
  array([ 1.74389351, -0.70860125, -1.58559951]),
  array([ 0.68526451, -0.25385001, -0.61849241]),
  array([ 1.95675841, -0.36534268,  0.22109845]),
  array([ 1.77372791, -0.480643  , -0.91669818]),
  array([ 0.74206509,  0.29681233, -0.26442804]),
  array([ 2.22242849, -0.27075169,  1.22332747]),
  array([ 2.52440456, -0.01480013,  1.26462558]),
  array([ 1.6097877 , -0.18556726,  0.74877403]),
  array([ 1.659671  , -1.72205302,  1.48431897]),
  array([ 1.89679487, -0.43862648,  1.12481435]),
  array([ 2.31669864,  0.18358242, -0.86063827]),
  array([ 0.82808131, -0.15406829, -1.0439539 ]),
  array([ 2.12752783, -1.71079394, -0.85409151]),
  array([ 1.48250585, -0.00944376,  1.34318311]),
  array([ 1.67381383,  0.68609953, -0.01952254]),
  array([ 0.81594867, -0.65449147, -0.66181601]),
  array([ 1.32571914, -0.09289073, -1.34219269]),
  array([ 0.5985487 , -0.50869366,  0.91366361]),
  array([ 1.15704437, -1.19276479, -0.72605884]),
  array([1.22182912, 1.31357981, 1.5902875 ]),
  array([ 0.51439707, -0.05833245, -0.38548623]),
  array([ 1.22579844, -0.24753921,  1.54837634]),
  array([1.30441103, 0.31510264, 1.71156889]),
  array([0.40273916, 0.3225213 , 0.68558742]),
  array([ 2.88545362, -1.33759476, -0.52409386]),
  array([ 0.65473773, -0.00522004,  1.72585464]),
  array([ 1.86151763, -0.56258684,  1.54741846]),
  array([0.34647862, 0.12324381, 0.21157687]),
  array([ 1.95147888,  0.39308172, -0.83834198]),
  array([2.33196259, 0.86867685, 1.53170908]),
  array([1.81043462, 0.49980689, 1.32435864]),
  array([1.16650782, 0.26746289, 1.54592282]),
  array([ 2.33201654,  0.27328796, -0.52654092]),
  array([ 0.6459248 , -1.33035002, -1.17356453]),
  array([ 1.715098  , -0.27844466,  1.50264783]),
  array([ 2.09869954, -0.94311696,  1.43000789]),
  array([1.29929009, 0.10366462, 1.04676267]),
  array([ 0.75643819, -0.47113539,  0.84378881]),
  array([ 2.24726283, -0.49998919,  0.1818356 ]),
  array([ 1.29284256,  0.01783367, -0.7321391 ]),
  array([ 0.45375142, -0.61036468,  0.87868061]),
  array([ 1.58024786, -1.41315256, -0.5764574 ]),
  array([1.61696442, 0.31376586, 1.17926237]),
  array([ 0.64679252, -0.02745916, -0.2878442 ]),
  array([1.31941503, 1.04603338, 0.69307319]),
  array([ 2.83459027, -0.944678  , -0.29372495]),
  array([ 1.08494528, -0.45300759, -0.4491951 ]),
  array([ 0.79628385,  0.39468976, -0.53888081]),
  array([0.95141331, 0.04870748, 0.78831631]),
  array([1.15981503, 0.52599894, 1.19814154]),
  array([ 1.18958919, -0.78401088,  0.42819372]),
  array([1.36218754, 0.37522961, 1.4824005 ]),
  array([ 1.44317408, -1.12647051,  0.83288979]),
  array([ 1.39205527, -1.67590088, -0.16442985]),
  array([ 2.16687981, -1.5532734 , -0.41088037]),
  array([ 0.86047408,  0.02659411, -0.77664794]),
  array([0.87390439, 0.13827263, 0.90147204]),
  array([ 0.71626505, -0.51572655, -0.58453346]),
  array([ 0.77250798, -0.99729011, -0.43230905]),
  array([ 1.63276965,  0.31973171, -1.13718237]),
  array([ 1.32008285, -0.01616106, -1.82957275]),
  array([ 2.45834026, -1.64118863, -0.42368886]),
  array([1.82689986, 2.09982405, 1.61262499]),
  array([ 1.04175701,  0.21523599, -0.4424665 ]),
  array([1.31730683, 0.64519731, 0.18826462]),
  array([ 2.37762381, -1.42195586,  1.1033111 ]),
  array([ 2.36046435, -1.2302806 ,  1.33326019]),
  array([1.92959685, 0.45523254, 1.28598457]),
  array([1.0814168 , 0.90511305, 1.62497827]),
  array([ 0.99124002, -0.51275285, -2.31612181]),
  array([ 1.38137598,  0.53833763, -0.18355434]),
  array([ 0.39049014,  0.07898617, -0.02962882]),
  array([1.67640526, 0.22944106, 0.99693549]),
  array([1.33048325, 1.44309527, 1.28088641]),
  array([ 0.41215134, -0.20942293,  1.17309198]),
  array([ 0.97505276, -0.62393434,  0.72887914]),
  array([ 1.14769321, -0.40411531,  1.28955857]),
  array([ 0.6999044 ,  0.25085716, -0.34643847]),
  array([0.63504575, 0.3918874 , 0.79757015]),
  array([ 1.12018297, -0.70036685, -1.04565759]),
  array([ 1.45710785, -0.29343927,  1.20818097]),
  array([ 2.76275359, -0.80886887, -0.29983097]),
  array([ 0.95102154, -0.31142868, -0.8380113 ]),
  array([ 2.22427018, -1.14488475, -0.37230784]),
  array([ 1.68845621, -1.22574954, -0.40132841]),
  array([ 2.46009403, -1.7769786 , -0.29470047]),
  array([ 1.99835951,  2.15002221, -0.20899507]),
  array([ 1.34591878, -0.29144564, -0.21245652]),
  array([ 0.69724764, -0.49926741, -0.33747668]),
  array([ 1.58175738, -1.73096942, -1.33406138]),
  array([ 1.67248445, -0.45005122,  1.53949758]),
  array([1.23051895, 0.49756083, 0.81940931]),
  array([2.92122674, 0.10165646, 0.66113205]),
  array([ 2.16595056, -1.06965112, -1.16721669]),
  array([1.15239334, 0.6131424 , 0.04396412]),
  array([ 2.60886034, -0.88849296, -0.39125574]),
  array([ 1.7654218 ,  0.97123091, -0.23508515]),
  array([ 0.92553574, -0.62645188,  1.37968709]),
  array([0.85286009, 0.72730575, 1.5762178 ]),
  array([ 1.34094628, -0.52632472,  1.41805198]),
  array([ 0.64772612, -0.03613725, -0.58082176]),
  array([ 2.35751543, -0.53306537,  0.99015951]),
  array([ 2.41587134, -0.72147192,  1.33468019]),
  array([ 1.70216227, -1.67786529,  0.94164447]),
  array([ 2.19141377, -0.0922409 ,  1.10979289]),
  array([ 2.30763977, -1.04285752,  0.80814939]),
  array([ 1.54237224, -0.43693883, -0.30758994]),
  array([ 0.67163778, -1.26824861, -0.56863085]),
  array([ 1.05829125, -0.50563897,  1.5738029 ]),
  array([2.63923903, 0.41078261, 0.04521532]),
  array([ 0.79412901, -0.1120304 , -0.29985425]),
  array([ 1.45288012, -1.3563036 , -0.06651155]),
  array([ 0.93932366, -0.83634732, -0.32325944]),
  array([ 2.2242385 , -1.19220447, -0.88307686]),
  array([ 1.38992548,  0.14730572, -1.1857636 ]),
  array([ 0.89934015, -0.62217924,  0.96533508]),
  array([ 1.08352613, -1.17717906, -0.45005058]),
  array([ 0.53169631, -0.58279392, -0.04115467]),
  array([ 2.57811217, -0.32183939,  1.08418484]),
  array([ 1.84594312, -1.75913166, -0.22866022]),
  array([1.1760453 , 0.33834607, 1.48121328]),
  array([0.32615518, 0.13808795, 1.76043904]),
  array([ 1.71148808, -1.113365  , -1.30289801]),
  array([1.67030679, 0.26137842, 1.46116638]),
  array([ 0.34722079, -0.27122875,  1.25284451]),
  array([ 0.80621302, -0.22349067,  0.44550569]),
  array([ 1.29071859, -1.14750954,  1.44621578]),
  array([ 1.93002255, -0.100701  ,  0.63998184]),
  array([0.93780127, 0.32580495, 0.498451  ]),
  array([ 1.46340933, -0.95765347, -1.92552029]),
  array([0.82405044, 0.44235905, 1.01727464]),
  array([ 2.11845601, -0.97296529,  0.89749434]),
  array([ 2.0291849 , -0.48233898,  1.5141546 ]),
  array([ 1.01032845, -0.06346028,  0.50964321]),
  array([ 2.00813933, -0.62907797,  0.0161249 ]),
  array([0.81030418, 0.85696438, 0.58800667]),
  array([ 0.75158824, -0.43880001,  1.62426343]),
  array([ 1.06148258, -0.60888606, -0.74798243]),
  array([ 0.36779748, -0.56193799,  1.26028195]),
  array([ 1.13091124, -0.76358228, -0.24882573]),
  array([ 1.78195239, -0.26449556, -0.48193004]),
  array([ 1.77777279, -1.68078102,  1.01840162]),
  array([ 1.4180489 , -0.86718116, -1.14179041]),
  array([ 0.78740792, -0.04013897, -0.54641007]),
  array([ 1.02578878, -0.54262145, -0.99355985]),
  array([ 2.60599518, -0.64757089,  1.44135132]),
  array([1.7168606 , 0.69404694, 1.29221973]),
  array([1.54967239, 0.87198558, 0.48242296]),
  array([ 2.02743727, -0.20018894,  0.89412659]),
  array([0.68281456, 0.0310174 , 1.56917311]),
  array([ 0.69263315, -0.26923439,  0.60327629]),
  array([ 1.6961548 , -0.63601163, -1.51366903]),
  array([ 0.76386835, -0.89098894,  0.34341918]),
  array([ 1.72607216,  0.39651516, -1.23978395]),
  array([1.09777186, 0.96684866, 1.02206688]),
  array([ 1.22188395, -1.26346624, -0.62981565]),
  array([0.86522543, 1.49273774, 1.75099384]),
  array([ 1.18351206, -1.49125095,  0.39266441]),
  array([ 0.40902595, -0.38322888, -0.46047418]),
  array([ 1.0242109 , -0.29689885, -0.74191167]),
  array([ 1.49102496, -1.65876075, -0.54267012]),
  array([ 0.66035384, -0.74319082, -0.65306147]),
  array([ 2.4431018 , -1.31959454,  1.20449967]),
  array([2.1977608 , 0.56925011, 0.91995664]),
  array([ 0.78930162, -0.27143909,  0.66961259]),
  array([ 0.87548334, -0.25811229, -0.29810317]),
  array([ 1.9634888 ,  0.28736306, -1.0454725 ]),
  array([ 1.24029497, -1.22387309, -0.25347995]),
  array([ 2.28361187, -1.47394063,  1.32324547]),
  array([ 0.72409622, -1.35746826,  0.86303115]),
  array([ 0.67655259, -0.65949465, -0.6815851 ]),
  array([ 0.41323064, -0.25983137, -0.20126673]),
  array([1.65053932, 0.37004674, 1.31823106]),
  array([ 0.8975503 , -1.43232688,  0.78971629]),
  array([ 1.49331456, -0.07359231, -0.31590748]),
  array([ 1.2879586 , -1.56609726, -1.18824132]),
  array([ 2.54522395, -0.62951478,  0.53517491]),
  array([ 2.17355505, -0.0826222 , -0.8490521 ]),
  array([1.90077576, 1.18972978, 1.41284961]),
  array([ 1.00301028, -1.64043853,  0.70702043]),
  array([ 0.65917474, -0.67392626, -0.72844398]),
  array([1.02453455, 1.20898101, 1.6996379 ]),
  array([ 1.56295826, -1.52875305, -2.82419956]),
  array([ 1.83210989, -1.45405802, -1.1475847 ]),
  array([ 2.08306199, -0.44155536,  0.50801361]),
  array([1.84664633, 0.42565749, 1.51772372]),
  array([ 1.77272505, -1.51501902,  1.41689526]),
  array([ 1.72143026, -0.42597128,  1.67092481]),
  array([0.96189857, 1.1574418 , 0.82883668]),
  array([ 1.95168656,  0.46225576, -0.38095027]),
  array([ 0.47594552, -0.65731045,  1.61959647]),
  array([1.30991743, 0.18193198, 0.89214974]),
  array([ 1.42980343, -0.95043051,  0.77378574]),
  array([ 1.56476657,  0.09359687, -1.85802479]),
  array([1.98176988, 0.43433391, 0.89005095]),
  array([ 1.92358943,  0.35788509, -0.21107956]),
  array([ 2.62896715, -1.67763287,  0.72138666]),
  array([ 1.49360528,  0.02729462, -0.37696303]),
  array([ 0.3892785 , -0.65443638,  1.06423752]),
  array([ 1.71733575, -1.13711143, -0.42033378]),
  array([ 0.79837359, -1.21536743, -0.34206992]),
  array([ 1.63032415, -0.76593058,  0.01557878]),
  array([ 1.57614405,  0.40662248, -2.56674802]),
  array([ 2.48230959, -0.87982759,  1.32946504]),
  array([1.35607992, 0.25570492, 1.02769786]),
  array([0.87648817, 0.26074765, 0.95916289]),
  array([ 1.63892995,  0.55280082, -0.46679916]),
  array([ 2.07530539, -0.18389242, -0.60525551]),
  array([0.69870556, 0.6525045 , 1.16237735]),
  array([ 1.28895659, -0.08442976, -1.18370657]),
  array([ 1.89041254, -0.81332114,  1.03248156]),
  array([ 2.07980557, -1.10375664,  0.9647834 ]),
  array([ 0.91967882, -0.11104888, -0.41186296]),
  array([ 1.56375326, -0.4198078 , -0.33454412]),
  array([2.16046781, 0.12069192, 1.42904449]),
  array([1.60556907, 1.38777623, 1.64847653]),
  array([ 1.89196725, -0.15456837, -0.33474576]),
  array([ 1.82946786, -0.13758426,  1.26237377]),
  array([ 0.97765232, -0.94635929, -0.34738892]),
  array([ 0.93727439, -1.49253808,  0.72064821]),
  array([ 1.92659566, -0.49624724,  0.00526178]),
  array([ 0.6929056 , -1.09513296,  1.64085684]),
  array([ 0.56023372, -0.02123541, -0.15202362]),
  array([ 2.61680294, -1.2779774 ,  0.3154009 ]),
  array([ 1.74519414, -0.05804074,  1.02827661]),
  array([ 2.80878339, -0.12459389, -1.25368341]),
  array([ 1.66511298, -0.39565651, -1.48131318]),
  array([ 2.85475691, -0.88766776, -0.00368516]),
  array([ 0.8551789 , -0.16062084, -0.17811633]),
  array([ 2.00531659, -0.80031117, -0.70119384]),
  array([ 1.05841932, -1.29406844, -0.12443368]),
  array([0.32538432, 0.49979432, 1.3703317 ]),
  array([ 2.28121051, -1.4790736 , -0.40370921]),
  array([2.02650091, 0.41293619, 0.36264086]),
  array([ 1.44758746,  0.52474124, -1.64745882]),
  array([ 1.15738592, -0.12166763, -1.15423641]),
  array([ 1.14258703, -0.04403797, -1.08963197]),
  array([ 0.76193044, -0.39746878, -0.31347525]),
  array([2.04601832, 0.29229553, 0.1337162 ]),
  array([ 2.2365158 , -1.1627127 , -0.40415148]),
  array([ 0.59918313, -0.09395826,  0.10605737]),
  array([ 0.89485629, -0.7536919 ,  1.44907529]),
  array([ 1.04892776, -0.63734832,  1.05788711]),
  array([ 0.42791582, -0.85753004,  1.47646714]),
  array([ 1.45563091, -0.35100323, -0.82130676]),
  array([ 0.37433151, -0.08221353,  0.11649719]),
  array([ 2.79499607, -0.49136991,  0.18895424]),
  array([ 1.76412144, -0.29818325, -2.69392219]),
  array([ 1.92578803, -1.22036784,  1.50163035]),
  array([0.88234025, 1.0130117 , 0.53222972]),
  array([ 1.55720791,  0.00695613, -0.27622251]),
  array([ 2.20301883, -0.85774282,  0.68019886]),
  array([ 0.9145653 ,  0.54206908, -0.14149164]),
  array([ 1.34191638, -0.09518693,  0.91610955]),
  array([ 1.29203588, -0.29340754,  1.55145008]),
  array([ 1.78598292, -1.1585716 , -0.44220462]),
  array([1.44048992, 0.37756489, 1.30863473]),
  array([1.57085549, 0.01016269, 0.50599707]),
  array([ 1.65967567, -0.59336561,  0.53940562]),
  array([0.28193864, 0.02574247, 0.78538322]),
  array([ 1.96631536,  0.63588827, -0.41046186]),
  array([ 0.366856  , -0.3025002 , -0.13790613]),
  array([1.95429549, 0.64177336, 1.32997924]),
  array([ 1.20882507, -0.13018777,  1.47659278]),
  array([1.26820983, 1.58279328, 1.50247379]),
  array([1.41864663, 0.93446314, 1.20379861]),
  array([ 1.72688625, -0.87243003, -0.34235374]),
  array([0.22379901, 0.17478466, 1.05063529]),
  array([ 1.41740886, -0.99760836, -0.32624586]),
  array([ 2.15083853, -0.5439837 , -1.2973887 ]),
  array([ 2.53567485, -1.62117924, -1.28049651]),
  array([ 1.26588401, -0.81590957,  0.41352621]),
  array([ 1.8170294 , -0.42932182,  1.19107224]),
  array([ 1.627933  , -0.30927311, -0.63655681]),
  array([ 1.65368108,  0.88317994, -0.22705483]),
  array([ 0.83183096, -1.06517303, -0.7806629 ]),
  array([ 1.71524519,  0.49143265, -0.32106046]),
  array([ 1.4446437 ,  1.07826464, -0.28094009]),
  array([ 1.39022406,  0.88337408, -0.45962932]),
  array([1.10060655, 0.10710184, 0.09440629]),
  array([ 1.39442694, -0.30120364, -0.24677676]),
  array([ 0.93590737, -0.18789759,  1.62127761]),
  array([2.54748047, 0.05072992, 1.1855675 ]),
  array([ 0.6671193, -0.6698876, -0.1674325]),
  array([ 8.74292932e-01, -1.76685908e+00, -1.73602510e-03]),
  array([2.33810325, 0.69140457, 1.53647142]),
  array([ 1.55222669, -0.56700251, -0.59212009]),
  array([ 1.90808918,  1.27300763, -0.39742276]),
  array([ 1.11585462, -0.53378177, -0.26364868]),
  array([ 1.78193445, -0.30290748, -1.04287166]),
  array([ 0.49372121, -0.160499  ,  0.13937629]),
  array([ 1.34378835, -0.77558609, -0.43752716]),
  array([ 2.03090037, -0.31195946,  0.6037509 ]),
  array([0.31012225, 0.62402327, 1.37867178]),
  array([ 0.8514405 , -1.07911765, -0.14887805]),
  array([ 1.1995925 , -1.13364389,  0.80205791]),
  array([ 1.12861113, -0.83722002,  1.66043078]),
  array([0.45942691, 0.26014286, 1.80546679]),
  array([2.39057589, 0.08393469, 1.33058461]),
  array([ 2.10713675, -1.27310721, -0.55876856]),
  array([ 0.87609591, -1.39858661, -0.20525959]),
  array([ 2.9520487 , -1.76642055, -0.49564082]),
  array([ 1.14040026, -0.46705797, -0.36509652]),
  array([ 1.17297777, -0.27992016,  1.00194857]),
  array([ 2.3208396 , -0.36939083,  1.46444224]),
  array([ 0.87942631, -0.71694102,  1.27801269]),
  array([0.40159695, 0.54512499, 1.68767627]),
  array([1.2650367 , 0.65081282, 0.68743978]),
  array([ 0.47862301, -0.98332192,  1.28293934]),
  array([ 1.44292029, -1.36621126,  0.69584152]),
  array([ 1.47610972, -0.28382957,  1.6838014 ]),
  array([ 0.54857272, -0.81379988, -1.22093559]),
  array([ 1.89687522, -0.43547431,  1.22892345]),
  array([ 1.83666668, -0.52054343,  1.29685068]),
  array([ 2.26526003, -0.84225085,  1.29570653]),
  array([ 1.22001645,  0.59870848, -0.36799213]),
  array([ 1.82625585, -0.56284483, -0.77635409]),
  array([ 0.45008246, -0.81467615, -0.11445169]),
  array([ 2.8065831 ,  0.42553713, -0.78387705]),
  array([0.40308648, 0.16933085, 1.50183464]),
  array([ 2.00550493, -0.27602454,  0.71045484]),
  array([ 0.64029006, -0.27606828,  1.1039185 ]),
  array([ 1.62269373,  0.31436405, -0.223895  ]),
  array([ 0.39521254, -0.35222352, -0.92401524]),
  array([ 1.49533607, -0.25999256, -0.59220865]),
  array([ 1.34802481, -1.15704414, -0.29554368]),
  array([ 2.43552765, -0.33686912,  1.12679236]),
  array([ 1.77317504, -0.14245258,  0.15148791]),
  array([0.80282484, 1.06794216, 1.21432721]),
  array([ 2.12245642, -0.57977735, -0.48936525]),
  array([ 0.91047472, -0.52448525, -0.45370745]),
  array([ 1.42917472, -0.56419122, -0.47667302]),
  array([2.22378454, 0.77382016, 0.55062987]),
  array([1.24916157, 0.15246553, 0.88770924]),
  array([ 2.3215017 , -0.55814527, -0.67308274]),
  array([ 0.82171495,  0.10668043, -0.23745768]),
  array([ 1.77693585, -1.77344723,  0.49783707]),
  array([1.44886833, 1.12899214, 1.41014767]),
  array([ 1.22307632, -0.18189713,  0.13857454]),
  array([ 0.65966322, -0.91762389,  1.60289379]),
  array([ 0.69190664,  0.19461228, -0.42946732]),
  array([ 0.63816947, -0.46859109, -0.99674544]),
  array([2.48171401, 0.07412402, 1.40081872]),
  array([ 2.2438335 , -0.16773793,  1.16046615]),
  array([ 0.6638702 , -0.19566573, -0.2754518 ]),
  array([ 0.51333162, -0.76638194,  1.23037673]),
  array([ 1.18774562, -0.31880232,  1.32251245]),
  array([ 2.34049226, -1.05840801, -1.17434009]),
  array([ 1.81692666, -0.61748494,  1.09677828]),
  array([ 1.14554731, -1.33020803,  0.25623417]),
  array([ 1.18016022,  0.19467877, -1.2699302 ]),
  array([ 1.02829969, -0.62164289, -1.41664088]),
  array([ 1.65580472, -1.74876614,  0.03075808]),
  array([ 0.84753961, -1.65974778, -0.74623001]),
  array([ 1.26188412, -0.46712302, -0.32477091]),
  array([ 1.05691842,  0.73052483, -0.41953913]),
  array([ 0.79058965,  0.49790952, -0.21394056]),
  array([ 0.81023377, -0.52150738,  1.49554997]),
  array([ 1.96775072, -0.56808874, -0.44203412]),
  array([ 1.90692073, -0.5080887 , -0.68237934]),
  array([ 1.52642266,  0.62681775, -0.33557148]),
  array([ 2.61972164, -0.90362802, -0.32694299]),
  array([ 1.4261407 , -0.44831628,  0.78390806]),
  array([1.38872106, 0.96423041, 1.44946442]),
  array([ 1.53727701, -0.42905943,  0.38768775]),
  array([ 1.39284798, -0.84377919,  1.40927299]),
  array([ 1.43145612, -0.31317853, -2.49484417]),
  array([ 1.31414715, -1.35658446,  1.40231766]),
  array([1.47524791, 0.1518576 , 0.75626864]),
  array([ 1.53351064, -0.29047853,  1.2373803 ]),
  array([ 2.4886118 , -0.62777879,  1.07187635]),
  array([0.20000081, 0.72956312, 1.7055121 ]),
  array([ 1.81934601, -0.49018048,  0.90945399]),
  array([1.33478467, 0.23730036, 0.29978721]),
  array([ 1.10647035, -0.43213431,  0.42595355]),
  array([ 0.97894927, -0.01635923,  1.54091623]),
  array([1.83417235, 0.13326712, 0.52318456]),
  array([ 0.82465911, -1.28473416,  0.14864558]),
  array([ 1.55191299,  0.37776063, -0.485577  ]),
  array([ 0.7361448 , -0.84275966,  1.26904921]),
  array([ 1.76891054, -0.67210689,  0.94572798]),
  array([0.87191628, 0.09371132, 0.76054662]),
  array([ 1.04146888, -0.62749668, -0.42486597]),
  array([ 1.2603326 , -0.66218588,  0.66314542]),
  array([2.43983729, 1.45679947, 1.21498793]),
  array([ 1.23713926, -1.24082808, -0.27537291]),
  array([1.63045152, 0.88771774, 0.85859192]),
  array([ 1.43207921, -1.22356128, -0.06561927]),
  array([0.55948935, 0.62502   , 1.10582363]),
  array([ 2.16349695, -1.25951727, -1.58828242]),
  array([ 1.83004329, -0.42780656,  0.81147919]),
  array([ 0.76041763, -1.49649427, -1.40832587]),
  array([0.29089313, 0.08958623, 0.55717696]),
  array([ 1.10409609, -1.06050928, -0.1423554 ]),
  array([ 1.37531191,  0.78139226, -0.58251887]),
  array([ 1.30267732,  0.50893877, -0.08322741]),
  array([0.90015369, 0.90425931, 1.23621706]),
  array([ 1.50832494, -1.66614372, -1.11148255]),
  array([ 1.61226325,  0.00163209, -0.85525443]),
  array([ 2.11827655, -0.07989226, -0.33629519]),
  array([ 2.2930414 , -1.01185562, -0.91179033]),
  array([0.58947411, 0.25505496, 1.04949094]),
  array([0.974213  , 0.74553769, 0.74862697]),
  array([1.68622211, 0.39960372, 1.58000024]),
  array([ 0.93146361, -1.58508245,  1.34540679]),
  array([ 1.03256835, -0.98895603,  0.40459206]),
  array([ 1.67456207, -1.58287617,  1.2918303 ]),
  array([ 1.76934168, -0.45527434,  1.34466951]),
  array([ 1.4855757 , -0.07405214,  1.14693149]),
  array([ 1.13114234, -0.77494052, -0.06129157]),
  array([ 1.88863127, -0.7922035 ,  1.43112987]),
  array([1.20210691, 0.0271143 , 0.815706  ]),
  array([ 0.50437476, -0.31732902,  0.1395261 ]),
  array([1.9146988 , 0.07430073, 0.6787805 ]),
  array([ 1.35414605,  0.32217965, -0.30086861]),
  array([ 2.11880947, -0.06532217,  0.9990606 ]),
  array([0.75717083, 0.2562399 , 0.21861142]),
  array([ 1.58811767, -0.4995274 ,  0.04746901]),
  array([0.26264328, 0.49960558, 1.4610931 ]),
  array([ 1.78105888, -1.1033546 ,  0.44616895]),
  array([ 0.68440675, -0.53515919, -0.27438231]),
  array([ 0.57844742, -0.86807112,  1.38684326]),
  array([ 1.42528196, -0.79261167, -0.23533448]),
  array([1.5034288 , 0.09352125, 0.71129299]),
  array([ 1.7432278 , -0.91920377,  0.28635939]),
  array([ 0.48797291, -0.53942423,  1.49711962]),
  array([ 1.33450548, -0.4251777 , -1.66097859]),
  array([ 2.39627871, -0.39760858,  0.61796431]),
  array([ 2.43921541,  0.30044592, -0.36610959]),
  array([1.38472276, 0.39751437, 0.3588784 ]),
  array([ 1.57631685, -0.32044081, -0.30106278]),
  array([ 0.79365053, -0.80431886, -0.63607351]),
  array([4.62088869e-01, 1.49133111e-03, 1.74664782e+00]),
  array([ 0.56621967, -0.42536785, -0.87366621]),
  array([ 0.91803172, -1.56935818, -0.32727507]),
  array([1.23417953, 0.3109875 , 1.59357386]),
  array([1.90441618, 0.35798219, 1.26913146]),
  array([ 2.1609746 ,  0.56748065, -1.18295064]),
  array([0.71844027, 0.25365433, 0.90761321]),
  array([0.81910746, 0.77469859, 1.4857312 ]),
  array([ 1.86156102,  0.47836309, -1.7545313 ]),
  array([ 1.97833486, -0.5819953 , -0.83387334]),
  array([ 1.91299622, -0.13549235,  1.52730417]),
  array([0.73647245, 0.3668643 , 0.14219829]),
  array([ 3.13713369, -1.11969869, -0.44650072]),
  array([ 1.93783026, -0.15339211,  1.23720927]),
  array([ 1.00369636, -1.78020882, -0.16134192]),
  array([0.66242988, 0.83508994, 1.34608241]),
  array([ 0.93914809, -1.45915451, -0.29478311]),
  array([ 0.79505022, -1.11487047, -0.29629335]),
  array([ 2.30974591, -0.52233119,  0.54629549]),
  array([ 0.76831174,  0.42248826, -0.18583715]),
  array([ 2.51827729, -0.70315091,  1.11056292]),
  array([ 1.32011013, -0.75487545, -0.7427311 ]),
  array([ 1.79206924, -1.03578531, -0.22760717]),
  array([ 1.1802384 , -1.18711703, -0.30141582]),
  array([ 1.3146346 ,  1.13288255, -0.2005271 ]),
  array([ 0.9290542 , -0.79617457, -2.45453644]),
  array([ 1.38173951,  0.35128878, -1.51666129]),
  array([ 2.16422545, -0.42794758, -0.07361001]),
  array([1.25412001, 1.63755963, 0.90676899]),
  array([ 1.38192654,  0.61460301, -0.27801416]),
  array([ 1.24238143, -0.72932984,  1.31470993]),
  array([ 2.26025653, -0.16496092,  1.28174895]),
  array([ 2.06201521, -1.73813233,  0.49450652]),
  array([ 1.36207637, -1.70327663,  1.44868925]),
  array([ 1.28419649, -0.3976243 , -0.9219667 ]),
  array([ 1.83467847, -0.63446094, -0.22588988]),
  array([0.94854808, 0.64553846, 1.20312297]),
  array([1.98706697, 1.41102436, 0.81414723]),
  array([0.70695198, 0.07844586, 1.57561878]),
  array([ 2.62931008, -1.24650302,  0.12644148]),
  array([ 1.94413706,  1.67520533, -0.31536455]),
  array([2.43029673, 1.21264177, 1.02294566]),
  array([1.43582061, 0.24375422, 0.26605543]),
  array([1.48036024, 1.61961456, 0.91684072]),
  array([0.45899436, 0.2165441 , 1.37247346]),
  array([ 0.86671712,  0.63857163, -0.18741374]),
  array([ 0.47931929, -0.05647183, -0.28139251]),
  array([1.26077115, 0.19643937, 1.59513432]),
  array([ 1.29161679, -0.52904278,  0.37038596]),
  array([0.9821303 , 0.22068424, 1.45048598]),
  array([ 1.37110263, -0.60689397,  1.48049624]),
  array([ 1.71104379, -0.34055068, -0.29584239]),
  array([ 0.45674443, -0.47151011,  1.6673308 ]),
  array([ 1.94280475,  0.02797718, -0.53709073]),
  array([ 0.36575826, -0.30899381,  1.28098713]),
  array([0.61232401, 0.37256488, 1.48016965]),
  array([ 0.86201165, -0.77080256, -1.86083508]),
  array([1.32066864, 1.59478617, 0.99290196]),
  array([ 1.81891508,  0.45713281, -1.29072922]),
  array([ 0.89021739, -1.69223176, -0.26325155]),
  array([ 2.53206176, -1.00348124,  0.19015479]),
  array([0.79727276, 0.322372  , 0.34342799]),
  array([ 2.20298973, -0.18352043, -1.13458211]),
  array([ 0.7580387 , -0.00854661, -0.23013859]),
  array([ 0.80555977, -0.68820223, -0.25347076]),
  array([ 0.55900352, -0.79551911,  0.01873346]),
  array([ 2.2647075 , -0.2934964 ,  0.74542301]),
  array([ 1.38883104, -0.81645894, -1.30574047]),
  array([1.48580906, 0.84008371, 0.33881068]),
  array([ 1.74083194, -0.61053965,  0.1637636 ]),
  array([0.56459243, 0.35727752, 0.57247922]),
  array([ 1.21700112,  0.24728072, -0.37975071]),
  array([ 0.38943458, -0.45781714, -0.46249324]),
  array([ 0.95671014, -1.12841418, -0.43820369]),
  array([ 1.96190131, -0.87395779, -0.33150653]),
  array([ 2.49354608, -1.52561418,  0.75517995]),
  array([ 1.67518868, -1.23629878, -1.85714285]),
  array([0.79704092, 0.22554036, 0.11183266]),
  array([ 0.7774047 , -1.20975344,  0.40907427]),
  array([ 2.24418561, -1.26471899,  1.44683334]),
  array([ 1.90679599, -1.51409893,  0.84019688]),
  array([0.69060564, 0.10856286, 0.7802061 ]),
  array([0.49995851, 0.54708246, 1.02443498]),
  array([1.32490362, 0.02670106, 1.07258596]),
  array([ 0.65911437, -1.13699299,  1.29674772]),
  array([ 1.7077637 , -1.45005704, -0.77001864]),
  array([ 2.17045658, -0.80041238, -0.78306398]),
  array([ 0.77697819, -0.67841003, -0.11182454]),
  array([1.74037998, 0.3162663 , 0.8005207 ]),
  array([ 0.428074  , -0.55019919,  0.11849032]),
  array([1.2010764 , 0.65599792, 1.37411423]),
  array([ 2.44525284, -0.43875779,  1.54093729]),
  array([ 0.87318592,  0.16831333, -0.34354666]),
  array([ 0.70626089, -0.47779386, -0.37948686]),
  array([ 1.31930671, -0.36840595, -0.99831596]),
  array([ 0.89640146, -1.4366721 ,  1.68440812]),
  array([ 2.3224531 , -0.57248758,  0.84498508]),
  array([ 0.81375858, -1.05515886, -0.27537567]),
  array([ 1.22783372, -0.75812635, -0.32675085]),
  array([ 2.34763242, -1.12255323,  1.03826478]),
  array([0.79889909, 0.34122203, 1.53799872]),
  array([1.47222038, 0.68610984, 1.50731503]),
  array([ 0.58890138, -1.64659046,  1.39152721]),
  array([ 1.89007666, -0.24801931,  0.70468221]),
  array([ 0.87545509,  0.07185478, -1.29291855]),
  array([ 1.61365516, -0.8343839 , -1.39992153]),
  array([ 0.57075208, -0.87032076, -1.41599454]),
  array([ 0.72534689, -0.13310888,  1.42816376]),
  array([ 0.80114797,  0.04230392, -0.48707389]),
  array([ 0.92704174, -0.81469191,  0.70955828]),
  array([ 1.00511505, -1.69241734, -0.40764167]),
  array([ 0.96388995,  0.20221334, -0.43378589]),
  array([ 1.82243184, -0.80774231,  1.44808431]),
  array([ 0.81900305, -1.52422668, -0.0735499 ]),
  array([ 1.66482756,  0.41081797, -0.24384899]),
  array([ 0.77402869, -0.33174165, -0.21182418]),
  array([ 0.49525894, -0.32724676,  1.53231732]),
  array([2.06942155, 0.40168803, 1.13722119]),
  array([ 2.64829693, -1.27565172,  1.3272132 ]),
  array([0.80693467, 0.52296768, 1.50742069]),
  array([0.64986996, 0.05396154, 1.5646054 ]),
  array([ 0.9997019 , -0.5891648 , -0.74161415]),
  array([0.23821439, 0.49598016, 1.26855313]),
  array([ 0.79025232, -1.18520281, -0.84320338]),
  array([ 0.63254766, -0.36497841, -1.06597844]),
  array([ 1.24147503, -0.5570619 ,  0.83991237]),
  array([ 2.25160234, -0.63921728,  0.18881297]),
  array([ 1.37290333, -0.34399879,  0.81204314]),
  array([ 1.92229804, -1.36378985,  0.74510686]),
  array([ 1.28822308, -1.14419518, -0.35620361]),
  array([1.36086262, 0.34074111, 0.50237199]),
  array([ 1.51126227, -0.2871047 ,  0.08609207]),
  array([ 2.56538498, -1.73223397,  1.28078503]),
  array([0.30160279, 0.02072245, 0.98332977]),
  array([0.35961682, 0.32969527, 0.76574181]),
  array([ 0.97780837, -1.09356967,  0.62286163]),
  array([ 0.50604653, -0.47571769, -0.11061166]),
  array([ 2.26074238, -1.06035072, -0.36564416]),
  array([ 2.28314016, -0.58477537,  0.73102944]),
  array([ 1.82270332, -1.12112486,  0.94572257]),
  array([0.70938321, 0.99299674, 1.17988085]),
  array([ 0.43505255, -0.12057295,  1.66755749]),
  array([ 2.06584223,  0.21101857, -0.1356784 ]),
  array([ 0.82652443, -0.21227432,  0.47963143]),
  array([2.42480007, 0.1598618 , 0.34681469]),
  array([ 0.88052053, -1.70077858, -0.70687888]),
  array([ 0.48097969, -0.09322892,  0.24197169]),
  array([ 2.41440594,  1.19457783, -1.18729762]),
  array([ 1.2189873 ,  0.17104038, -0.42128212]),
  array([ 1.55487896,  1.02171106, -0.45326192]),
  array([ 1.20264633,  0.72328473, -0.25922705]),
  array([ 0.95339913, -0.81562322, -0.3219527 ]),
  array([ 1.09648722, -1.48997675, -1.06439903]),
  array([ 1.83441502, -0.06168958, -2.44570924]),
  array([ 0.51804002, -0.1989758 ,  1.09326586]),
  array([ 1.24957159, -1.4215672 ,  1.1866802 ]),
  array([1.97472355, 0.21852239, 0.73160082]),
  array([ 1.1666223 , -0.21855868,  1.30675087]),
  array([ 2.77984951, -0.56453514,  0.74813991]),
  array([ 1.00516137,  0.39176809, -0.12746897]),
  array([ 0.48818137,  0.10075082, -0.23760564]),
  array([ 2.03815673, -0.21776668,  0.855642  ]),
  array([ 1.88466265, -1.19617442, -0.60880787]),
  array([ 0.80111651, -0.68902538,  1.5542878 ]),
  array([1.3890769 , 0.55679335, 1.42818215]),
  array([0.34593935, 0.08004602, 0.89001295]),
  array([ 0.69274013, -0.10205669, -0.25148346]),
  array([ 1.39962401, -0.05963486, -0.06989752]),
  array([ 0.87812495,  0.44381163, -0.2686786 ]),
  array([ 0.59974565, -0.79422097,  0.7966743 ]),
  array([ 2.47457247, -0.30172003, -0.48121795]),
  array([1.17275551, 0.65383947, 1.32808818]),
  array([ 1.79786083, -0.79091698, -0.35597063]),
  array([ 0.70054232, -0.86941536, -0.14229084]),
  array([ 1.93896104, -0.39581577, -0.8985896 ]),
  array([ 1.63480168,  0.0192634 , -1.14236309]),
  array([ 1.85407111, -0.53378984,  0.03817429]),
  array([1.25919053, 1.9798383 , 1.40264942]),
  array([ 2.64397071, -0.27530072,  1.25073081]),
  array([ 1.82491814, -0.1136677 ,  0.30477038]),
  array([0.55134977, 0.58488873, 0.74753242]),
  array([ 1.93387946, -0.145058  ,  0.69449522]),
  array([ 1.02094319,  0.54241236, -0.2775597 ]),
  array([0.4201706 , 0.30518574, 0.87615914]),
  array([2.89784894, 0.57033801, 1.36520656]),
  array([0.58611765, 0.77421113, 1.03743123]),
  array([ 0.80471643, -0.44119198,  0.20661276]),
  array([ 1.48308471,  0.18527505, -0.56227237]),
  array([ 0.81079227, -0.0650897 , -0.97061934]),
  array([ 1.44876468, -0.25341851,  0.05995329]),
  array([ 1.30780902, -0.6589904 ,  1.49831852]),
  array([ 0.65534981, -0.10970316, -0.52237113]),
  array([1.33098159, 1.1574392 , 1.42725316]),
  array([ 1.28296134,  0.06304227, -0.27373299]),
  array([1.18572448, 0.3388196 , 0.84667889]),
  array([ 0.82653427, -0.14704103,  0.95397557]),
  array([ 1.72577447, -1.38199573, -0.38190011]),
  array([0.22332647, 0.13987959, 1.01872578]),
  array([ 1.13947272,  0.58602051, -0.31908791]),
  array([ 1.25590429, -0.21934829,  0.78818086]),
  array([ 1.92973472, -0.58115721,  1.52233149]),
  array([1.75218633, 1.36568406, 0.56620963]),
  array([ 0.83872262, -0.50925637, -0.41028645]),
  array([ 3.26246828,  0.87971982, -1.70074306]),
  array([ 1.69317838, -0.39123852, -0.39596404]),
  array([ 1.60765691,  0.50173416, -0.36571154]),
  array([ 1.87391537, -1.11012118,  1.21826996]),
  array([2.59780806, 1.2975642 , 0.63067382]),
  array([ 2.62033787, -1.29414388,  0.56624968]),
  array([ 0.83910719, -0.63252168,  0.17733754]),
  array([ 2.10075115, -1.4180648 , -0.40043929]),
  array([ 1.32141303, -0.62513159, -0.39267267]),
  array([ 1.64100072, -0.84508913,  1.30025462]),
  array([ 1.92627761, -0.78214537,  0.96090914]),
  array([ 1.66596657, -0.91980409,  1.37922639]),
  array([ 1.68088972, -1.19138257,  0.39625166]),
  array([2.22518241, 0.28803282, 1.01647345]),
  array([ 2.06491732, -0.94696201, -0.50043497]),
  array([ 0.89908203, -0.16897689, -0.97847277]),
  array([ 1.30614085, -0.76925284,  0.70010626]),
  array([ 1.06076872, -0.42857196,  1.57099848]),
  array([ 2.65561977, -1.15646218,  1.38869493]),
  array([ 1.43567807,  0.7454841 , -0.25019357]),
  array([1.55678085, 0.73021144, 0.94750831]),
  array([ 1.89924088, -1.68628013, -0.35834619]),
  array([ 2.10689166, -0.40479576,  0.01584136]),
  array([ 0.80951288, -0.62830499, -0.06434822]),
  array([ 0.81053759, -0.25725262, -1.03025789]),
  array([0.90075331, 0.03926972, 0.81305177]),
  array([ 1.34844318, -1.36129136, -0.24007088]),
  array([ 0.68155191, -0.45297562, -0.40545199]),
  array([ 0.60553396, -0.51618516,  1.64684535]),
  array([1.02903979, 0.40786464, 0.88535354]),
  array([0.80520717, 1.34318925, 1.80182048]),
  array([0.62278352, 0.81829426, 1.81904884]),
  array([1.35132799, 0.44032009, 1.62814158]),
  array([ 1.8635938 , -0.86525334, -0.27624828]),
  array([1.11656462, 1.30176302, 0.89409642]),
  array([ 1.55867768, -0.98249411,  0.24595185]),
  array([ 2.29682932,  0.78727501, -1.24069357]),
  array([ 1.38002966, -0.19031034, -0.18991667]),
  array([ 1.24010276,  0.39131548, -0.22087671]),
  array([ 1.65993837, -0.72218937, -0.15676755]),
  array([ 1.08418208, -0.66979378,  0.21817091]),
  array([0.55510909, 0.62050087, 1.79957417]),
  array([ 2.26128079, -0.82665306, -0.63782018]),
  array([2.01747654, 0.54662909, 0.40207172]),
  array([ 1.24171069, -0.55600871, -1.21127899]),
  array([2.18536927, 1.78316805, 0.25724117]),
  array([ 2.1248667 ,  0.39902462, -0.53971134]),
  array([ 1.47176601, -0.28631795, -0.46353011]),
  array([ 1.01704761, -0.39213753, -0.56816743]),
  array([ 0.41845097, -0.52238072, -0.22441436]),
  array([ 2.34010584, -1.08504896, -0.42544335]),
  array([ 2.29860397, -0.89640304, -2.34959303]),
  array([ 2.05904381, -0.71612853, -0.79589565]),
  array([ 0.70152182, -0.3890892 , -0.47166653]),
  array([ 1.91982999, -0.46927751,  1.32776558]),
  array([ 2.79504553, -1.10925482,  0.59096671]),
  array([ 0.93368466, -0.93963342, -0.27193787]),
  array([ 2.09866746, -0.62100122, -0.30312046]),
  array([0.67129928, 0.97582273, 1.29725337]),
  array([ 2.2631433 , -0.9631574 , -0.58193444]),
  array([0.96626125, 0.86803464, 0.91833233]),
  array([ 2.32542417, -0.35104296,  0.46822395]),
  array([ 0.81733668, -1.66169512,  1.40091647]),
  array([1.57364567, 0.52000947, 0.41042562]),
  array([0.46735416, 0.01100273, 1.24048734]),
  array([ 1.1215162 ,  0.40622882, -0.79868414]),
  array([ 1.28188398, -0.52456982,  0.87683489]),
  array([ 1.56070154,  0.23933355, -0.38270255]),
  array([ 2.53135473,  0.80312935, -0.58890359]),
  array([ 1.09252184, -0.22704129,  0.34247751]),
  array([ 1.63002723, -1.05596419, -0.56521748]),
  array([ 1.64391252, -1.51735624, -0.69023374]),
  array([ 1.16569997, -0.33749907,  0.8527963 ]),
  array([ 1.78824832,  0.88850278, -0.02921353]),
  array([ 0.42791531, -0.14527317,  1.59559991]),
  array([1.39692591, 0.1528888 , 1.60221346]),
  array([1.35903083, 0.75988726, 0.7972881 ]),
  array([ 1.10856823, -0.5598098 ,  1.53532237]),
  array([ 2.6563777 , -0.78062577, -0.88278906]),
  array([ 0.67806879, -0.23690566, -1.10299799]),
  array([ 1.90172885, -0.42542928, -0.18275766]),
  array([ 1.64319425,  1.70284926, -0.22729703]),
  array([ 1.59846707, -0.71594937,  1.51448824]),
  array([0.87193427, 0.24729045, 1.26743075]),
  array([ 0.73484094, -0.92186787, -2.04249169]),
  array([ 0.73138061, -0.47848147,  1.35566019]),
  array([ 0.90511888, -0.43760191,  0.91363902]),
  array([ 0.62364936,  0.13152027, -0.21095648]),
  array([1.15393591, 0.82021148, 1.46361484]),
  array([ 2.16101236,  1.48293686, -0.53612698]),
  array([ 0.4708894 , -0.11541735,  0.69200895]),
  array([ 1.16959521, -1.17261497, -0.56296803]),
  array([ 1.61439353,  0.66243029, -1.10387394]),
  array([ 2.0803598 , -0.18108152,  1.4305233 ]),
  array([1.77429893, 0.54833485, 1.34111333]),
  array([1.66905959, 0.42373045, 0.40892746]),
  array([0.29745882, 0.24670828, 1.07041138]),
  array([ 0.38580008, -0.02300596,  1.55789312]),
  array([1.64261688, 0.32524352, 0.89597612]),
  array([ 1.5480624 , -0.30773666,  0.73792071]),
  array([ 0.80057305, -0.64379602,  0.11276228]),
  array([ 2.72492476, -0.87274253, -0.34562587]),
  array([1.40392707, 0.06918596, 1.23653381]),
  array([ 1.5242617 ,  0.07103557, -1.00950607]),
  array([ 0.64177429,  0.24736083, -0.14164433]),
  array([1.02865127, 0.15685386, 0.17580364]),
  array([ 2.72435648, -1.4671826 ,  0.09213958]),
  array([ 0.92600745, -1.05112106,  0.34570868]),
  array([ 0.38024877, -0.04403851,  0.5856886 ]),
  array([0.46813553, 0.95070143, 1.78010258]),
  array([ 2.04455796, -1.72205539,  1.45669934]),
  array([ 2.03738235, -0.36170696, -0.53228196]),
  array([ 1.09523   , -0.30553645,  0.51625331]),
  array([ 1.32358394, -0.13859481,  0.38296298]),
  array([ 0.84597994, -0.5439969 , -0.78267611]),
  array([ 2.57702113, -0.58256412,  0.1516468 ]),
  array([ 1.54791075, -0.43236845, -0.29658778]),
  array([0.77018943, 0.69626226, 1.0373364 ]),
  array([2.02895911, 0.16463816, 0.77072412]),
  array([ 2.42655622, -0.81399994,  0.07305422]),
  array([ 1.08369547, -0.40003569,  1.12330674]),
  array([ 0.97597447,  0.65328531, -0.49030031]),
  array([1.17687532, 0.33678167, 0.77098356]),
  array([ 0.70584547, -0.39951717, -0.33935478]),
  array([ 2.06153415, -0.57726287,  0.98116222]),
  array([ 2.42191044, -1.34759156, -0.43940725]),
  array([ 0.46485742, -0.42547544, -0.12254513]),
  array([ 0.66698471, -0.76165978, -1.17550886]),
  array([ 0.80982618,  0.11927708, -0.52537595]),
  array([ 0.6842905 , -0.24114587, -0.13959331]),
  array([ 1.58796447, -0.13803539,  0.30702079]),
  array([1.00605346, 0.15632054, 1.57135618]),
  array([ 1.74149473, -0.49283905,  0.39010382]),
  array([0.78205575, 0.31678406, 1.0223505 ]),
  array([ 1.6270604 , -0.1671879 ,  0.20367742]),
  array([ 1.28537938,  0.28211881, -0.1370715 ]),
  array([ 1.29711019, -0.366701  , -0.65118219]),
  array([ 1.00297107, -0.66822032,  0.67293583]),
  array([ 0.89206296, -0.95966237, -1.88136732]),
  array([ 1.25665567,  0.82434107, -0.09280084]),
  array([ 2.12565896,  0.20402978, -0.38483516]),
  array([ 0.5186698 , -0.30687205, -1.05160769]),
  array([ 0.6246218 , -1.03868689, -0.28805323]),
  array([ 0.90948053, -0.83601311, -0.44074975]),
  array([ 2.00404724, -0.58013588,  1.48437696]),
  array([ 1.09741081, -1.29566493,  0.41263173]),
  array([ 0.84627594, -0.2071597 , -1.3706388 ]),
  array([ 1.86876394, -1.75386704,  0.62512097]),
  array([1.3174921 , 1.34549257, 1.58066499]),
  array([1.65440314, 1.34457211, 1.48900369]),
  array([ 0.31566966, -0.09594056,  0.51110141]),
  array([ 2.9397482 , -1.30365269,  0.27537148]),
  array([ 2.45486577, -1.29991575, -0.62612578]),
  array([ 1.64796981, -0.43548604,  0.31239574]),
  array([ 0.85920283, -0.8609404 ,  0.86768548]),
  array([1.29201249, 0.497229  , 0.47247348]),
  array([ 1.71679191, -1.65385342,  0.64428502]),
  array([ 0.6941002 , -0.69794425,  1.18083953]),
  array([2.32412619, 0.06496516, 0.61157884]),
  array([1.79858024, 0.92843085, 1.17946476]),
  array([1.83995118, 1.01514242, 0.22108875]),
  array([ 1.91628954, -0.78044048,  0.99911973]),
  array([1.61233072, 1.3636038 , 1.4758075 ]),
  array([ 0.4874254 , -0.92545795,  1.66771555]),
  array([ 0.60787385, -0.64131274, -0.07996821]),
  array([0.9975795 , 0.18771763, 0.54985089]),
  array([ 1.29720281, -0.46384304,  0.61715134]),
  array([ 0.93462173, -1.28123398,  0.17872494]),
  array([ 1.47154501,  0.34436005, -0.23292299]),
  array([ 1.12263   ,  0.4932395 , -0.20743435]),
  array([ 2.69909479, -1.17575822,  0.24314416]),
  array([ 2.24132537, -0.36437749,  0.54850429]),
  array([0.21600208, 0.22918701, 1.38234483]),
  array([ 1.09634905,  0.36686807, -0.66910224]),
  array([ 1.4204976 , -1.65572982, -0.86450247]),
  array([2.38083627, 0.03144083, 1.14948944]),
  array([ 1.63400122,  1.30698358, -0.45661151]),
  array([ 0.61547276, -0.66252855,  1.15120578]),
  array([0.88300512, 0.86856749, 0.61302487]),
  array([ 2.59477052, -0.60520034, -0.49202173]),
  array([ 1.31728636, -0.60813203, -0.3367777 ]),
  array([1.03356284, 1.13849409, 1.59301188]),
  array([ 2.24741938, -0.33577624,  1.54208768]),
  array([ 1.60984576, -0.38027759,  1.44254393]),
  array([0.47186493, 0.3044309 , 0.40918851]),
  array([1.77021275, 1.47501549, 1.29111574]),
  array([ 0.40427272, -0.23993408, -0.19912196]),
  array([ 1.05425251, -1.18609888, -0.26103601]),
  array([ 2.41580706, -1.23209275, -0.01752154]),
  array([ 1.49439621, -0.48621134, -0.4300369 ]),
  array([ 1.60870433e+00,  2.66906810e-04, -5.94310054e-02]),
  array([1.13888972, 0.17056064, 1.62240545]),
  array([ 1.05285298, -1.44581094, -0.29296513]),
  array([ 2.79561443, -1.13265454,  1.04003956]),
  array([ 1.73310601, -0.8452729 , -0.5790398 ]),
  array([ 1.38446524,  0.25938465, -1.14433343]),
  array([ 0.77411303, -1.35686585, -1.73233044]),
  array([ 1.28685897, -0.43974901, -0.62096531]),
  array([ 0.87473713, -0.2878169 , -0.46612703]),
  array([ 0.47202561, -0.77882778, -0.45021327]),
  array([1.65346558, 0.76714359, 0.82336272]),
  array([ 1.60183611, -0.39424128,  0.93886059]),
  array([ 1.26228214, -0.61220429,  0.6058675 ]),
  array([ 1.3085266 , -0.80771571,  1.4975984 ]),
  array([ 1.09129789,  0.40072194, -0.26082972]),
  array([1.44424825, 0.16655114, 1.35094234]),
  array([ 0.62222235, -0.30325198, -0.34391869]),
  array([ 2.1938922 , -0.83188692,  0.68371708]),
  array([0.55670727, 0.95155792, 1.16849835]),
  array([1.68237734, 2.43279141, 1.51752982]),
  array([ 0.92356876, -1.14962936,  0.71174055]),
  array([ 0.98749238, -0.63994445, -0.25970276]),
  array([ 2.08283373, -0.48443001,  0.109147  ]),
  array([1.2578667 , 1.11530114, 0.16082547]),
  array([ 2.26430447, -0.32611315, -3.24041537]),
  array([ 2.29972376, -0.72306403,  0.67861731]),
  array([ 1.29950256,  0.18520634, -0.42056625]),
  array([ 0.93820881, -0.77251314, -1.20290728]),
  array([ 0.92197433, -0.48275038,  0.97716944]),
  array([1.20320921, 0.88117388, 1.41287006]),
  array([ 0.73628244, -0.37074816,  1.36037258]),
  array([0.25608518, 0.32532344, 1.21989973]),
  array([ 0.89956545, -0.14420855, -0.27836943]),
  array([ 1.46666151, -1.47681835, -0.36209547]),
  array([ 0.67819498, -1.41545972, -0.37013802]),
  array([0.59987845, 0.19916365, 0.34426917]),
  array([ 0.505975  , -0.24704258, -0.21396203]),
  array([2.05160579, 0.12637165, 0.73097147]),
  array([ 1.35387396, -0.13096319, -1.18823811]),
  array([ 2.03611855, -1.45534238, -0.32464218]),
  array([ 1.89053925, -0.44058208, -1.4722876 ]),
  array([ 2.6337501 , -1.20210554,  0.99734569]),
  array([ 1.79223322, -0.31422356, -0.73587   ]),
  array([ 0.5481003 , -0.48974811, -0.79105131]),
  array([1.36680064, 1.0147652 , 1.74851016]),
  array([1.20839546, 0.40376522, 0.12925406]),
  array([ 0.70722221, -0.02887359, -0.09215723]),
  array([ 1.54997469, -0.22435969, -0.181694  ]),
  array([1.12865044, 1.09807764, 0.05379626]),
  array([ 2.26818464,  0.82911235, -0.74146758]),
  array([ 1.41963833, -0.08698354, -1.12412781]),
  array([ 1.07594309, -0.6400179 ,  1.40984625]),
  array([ 1.87765275, -0.21247427,  0.60117833]),
  array([ 1.41676392, -0.49982664, -1.71795489]),
  array([0.84845728, 0.07799505, 1.14214528]),
  array([ 0.56205351, -0.45698598, -1.27005327]),
  array([ 1.37251249,  0.04194056, -0.2496203 ]),
  array([ 1.98218102, -1.30692168, -1.51560662]),
  array([2.07334252, 0.07146818, 1.26203584]),
  array([ 1.81619439,  0.07604277, -1.10842688]),
  array([ 0.85175638, -0.38282052,  1.02047065]),
  array([ 1.48078279, -0.30250138,  1.14732643]),
  array([ 2.74089264,  0.26870446, -0.61441936]),
  array([ 2.91418965, -0.9466952 ,  0.75170492]),
  array([ 0.70907585, -0.0914709 ,  0.3650305 ]),
  array([ 1.8473649 , -1.27479938,  0.10994421]),
  array([ 1.46055715, -0.32390191,  1.05555549]),
  array([ 3.05283513, -0.20011789, -1.02578118]),
  array([ 1.76367268, -1.48995059,  1.51644865]),
  array([0.45977323, 0.24483277, 1.2206102 ]),
  array([ 0.55975361, -0.39401403, -0.32093434]),
  array([2.22206712, 0.13137026, 1.44348171]),
  array([ 0.36161128, -0.53293091,  1.29354071]),
  array([0.75002555, 0.49040636, 0.57347056]),
  array([ 2.64237979, -1.02160452, -0.20096621]),
  array([ 2.07470457, -0.97213472,  1.26627924]),
  array([0.98540102, 0.28052832, 0.65709591]),
  array([1.40898633, 0.58423096, 0.14727004]),
  array([ 1.49050039, -0.68821774,  1.1363791 ]),
  array([ 1.26827098, -0.75672912,  1.58096019]),
  array([ 1.47642298, -0.14581576,  0.32597293]),
  array([1.63075686, 0.12465821, 0.22554437]),
  array([ 1.86290846, -1.17330395,  1.08425626]),
  array([0.86124892, 0.31035004, 1.19090348]),
  array([2.12336079, 0.6632336 , 0.66894481]),
  array([ 2.52201387,  0.4950925 , -0.99584699]),
  array([ 2.16826327, -1.23200999, -0.09799268]),
  array([ 1.40976826, -1.06691508,  0.67472677]),
  array([ 1.43824638, -0.25634795,  0.13006093]),
  array([ 1.96555948,  0.21138446, -0.05718831]),
  array([ 0.86195777, -0.91252451, -0.37099517]),
  array([ 1.30230735, -0.18173633, -1.00052814]),
  array([ 2.02174483, -0.38296544,  0.8215716 ]),
  array([ 2.13473447, -0.0226218 ,  0.66360751]),
  array([ 2.23550682, -0.72061295, -1.32441752]),
  array([ 0.69064117, -1.01320939, -1.17900615]),
  array([ 1.63700597, -1.4713817 , -1.04782877]),
  array([2.21703639, 0.67264393, 0.62335918]),
  array([ 0.84385698, -0.52934959, -0.29237179]),
  array([ 0.64520307, -0.60759306,  1.2228681 ]),
  array([ 1.48206908,  0.57483235, -0.59935223]),
  array([ 1.96069738, -0.17924342,  1.03499307]),
  array([ 0.75862822, -0.78522168,  1.19409976]),
  array([ 1.68744884, -0.26849105, -0.37587836]),
  array([2.37655486, 1.35955206, 0.34613442]),
  array([ 1.51705718, -0.48938135,  1.60722928]),
  array([ 0.43732262, -0.42211598,  1.43216758]),
  array([ 0.48142985, -0.14299834, -0.17458188]),
  array([0.79009224, 0.5199831 , 1.22097695]),
  array([1.83249656, 0.07555012, 0.29627794]),
  array([0.75097438, 0.13931679, 0.60875974]),
  array([ 0.62346244,  0.1352521 , -0.56214731]),
  array([2.02789681, 0.29918054, 0.90203403]),
  array([ 0.99291068, -1.78234677, -2.66595913]),
  array([ 1.20353886,  0.37584222, -0.025919  ]),
  array([ 0.85337957, -0.17837147,  0.31659339]),
  array([ 1.98766894, -0.36390588, -0.52677599]),
  array([ 0.7885776 , -1.25956745,  1.45119095]),
  array([ 2.82594116, -0.42435467, -0.39171951]),
  array([ 1.50502479, -0.48259798,  0.51849417]),
  array([1.206483  , 0.81533441, 1.27881574]),
  array([ 0.77609296, -1.59233132,  0.17365936]),
  array([ 1.33711651, -0.52143679,  0.38976966]),
  array([ 2.04239305, -1.71016375, -0.48869852]),
  array([ 2.84969971, -1.72650481,  1.36055409]),
  array([1.70613278, 1.50420709, 0.660499  ]),
  array([ 0.6476718 , -0.99095056, -0.23148492]),
  array([ 1.07820907, -0.61417601,  0.92894185]),
  array([ 0.68066722, -0.41736211, -1.03913152]),
  array([0.34176089, 0.14928708, 1.16270374]),
  array([ 1.03744289, -0.80712203,  1.68059197]),
  array([ 1.63925113,  0.5650751 , -1.21855652]),
  array([ 1.66801162,  0.31066778, -0.79758095]),
  array([1.01307548, 0.17393492, 1.47427845]),
  array([ 0.57486994, -1.21785966,  1.34335384]),
  array([ 1.45328658,  0.15386581, -1.19329157]),
  ...]]

# define two function that the cell needs
def get_dis_sta(df):
    #获得某个cluster的距离标准差的最大分量
    sta = []
    for i in range(df.shape[1]):
        #print(df.iloc[:,i])
        sta.append(df.iloc[:,i].std())
    x = max(sta)
    return x

def get_cen_dis(list_cen):
    #计算所有簇的中心点间的两两的距离
    dis_cen = []
    for i in range(len(list_cen) - 1):
        for j in range(i+1 , len(list_cen)):
            dis_cen.append(get_distance(list_cen[i] , list_cen[j]))
    return dis_cen

# show the result of the kmeans clustering, and get the parameters that ISODATA clustering needs
plt.rcParams['figure.dpi'] = 150 #分辨率
cluster_num = []   #记录各个簇的样本数量
cluster_sta = []   #记录各个簇的标准差最大分量
cluster_cen = []   #记录各个簇的簇中心点

fig = plt.figure()
ax = Axes3D(fig)
ax.set_xlabel("Factor 1") ; ax.set_ylabel("Factor 2") ; ax.set_zlabel("Factor 3") 

for cluster in result:
    df_clu = pd.DataFrame(cluster)
    df_clu.rename(columns = {0 : 'f1' , 1 : 'f2' , 2 : 'f3'} , inplace = True)
    cluster_num.append(df_clu.shape[0])
    cluster_sta.append(get_dis_sta(df_clu))
    cluster_cen.append([df_clu.mean()[0] , df_clu.mean()[1] , df_clu.mean()[2]])
    ax.scatter(df_clu['f1'] , df_clu['f2'] , df_clu['f3'] , s = 1, c=None )
    
dis_cen = get_cen_dis(cluster_cen)
print("各个簇的样本个数:" , cluster_num)
print("各个簇的标准差最大分量:" , cluster_sta)
#print("各个簇的中心坐标值:" , cluster_cen)
print("各个簇中心间的距离:" , dis_cen)
plt.show()

各个簇的样本个数: [1632, 1812, 1179]
各个簇的标准差最大分量: [1.00419530633098, 0.932311982173369, 0.9450386081567261]
各个簇中心间的距离: [1.7647008995850917, 2.1417655652416903, 2.116485100906183]

# get the kmeans's result's Silhouette Coefficient
df_kmeans = pd.DataFrame(columns = ['f1' , 'f2' , 'f3' , 'KMEANS']) 
for i in range(len(result)):
    df_temp = pd.DataFrame(result[i])
    df_temp.rename(columns = {0 : 'f1' , 1 : 'f2' , 2 : 'f3'} , inplace = True)
    df_temp['KMEANS'] = i
    df_kmeans = df_kmeans.append(df_temp)
df_kmeans.reset_index(drop = True , inplace = True)
# 实际上下面这种表联结的方式存在一些问题，因为如果有几辆车的factor1 , factor2 , factor3完全相同的话，
# 那么就会出现他们彼此间两两都会联结的情况了，这就导致样本集的数量增多了，但是这种情况是可以接受的，
# 因为第一出现这种情况的概率是很小的，第二如果两辆车的feature都相同，则必定会被聚到一个类，直接去重是不会影响的。
df_kmeans = pd.merge(df_feat_new , df_kmeans , on = ['f1' , 'f2' , 'f3'] , how = 'inner')
df_kmeans.drop_duplicates(subset=['CAR_NUM'], keep='first', inplace = True)
df_kmeans['KMEANS'] = df_kmeans['KMEANS'].astype(np.int)  # 将ISODATA这一列转化为np.int类型，使得describe()可以成功对这一列做统计
df_kmeans.head(5)

	f1	f2	f3	CAR_NUM	DBSCAN	F1	F2	O1	D1	O2	D2	FS	STD_FD	AVG_FD	KMEANS
0	0.633961	0.057235	-1.165552	浙AFZ391	1	0.047619	0.904762	0.142857	0.142857	0.571429	0.619048	-2.000000	2.093641	5.125000	0
1	-0.878013	-0.763536	-0.307468	浙B732UE	0	0.047619	0.428571	0.047619	0.047619	0.047619	0.047619	-1.453860	1.118034	2.333333	1
2	1.312621	0.156470	0.043351	浙AN0G37	-1	0.428571	0.952381	0.523810	0.523810	0.523810	0.523810	-0.677038	2.547726	4.909091	2
3	-0.335299	-0.023774	-0.200037	浙A3WA22	0	0.047619	0.571429	0.047619	0.047619	0.285714	0.333333	-1.199703	2.958040	3.666667	1
4	0.847105	1.263833	-0.014173	浙GN820X	-1	0.476190	0.857143	0.476190	0.285714	0.285714	0.523810	-0.820671	5.780846	5.727273	0

# get the kmeans's result's Silhouette Coefficient
silhouettevalue = metrics.silhouette_score(df_kmeans[['f1' , 'f2' , 'f3']].values , df_kmeans['KMEANS'] , metric='euclidean')
silhouettevalue

0.2757943074884164

Understanding KMEANS clustering's statistical result

# get the first cluster's statistical result
df_kmeans[df_kmeans['KMEANS'] == 0][['F1' , 'F2' , 'O1' , 'D1' , 'O2' , 'D2' , 'FS' , 'STD_FD' , 'AVG_FD' , 'KMEANS']].describe()

	F1	F2	O1	D1	O2	D2	FS	STD_FD	AVG_FD	KMEANS
count	1632.000000	1632.000000	1632.000000	1632.000000	1632.000000	1632.000000	1632.000000	1632.000000	1632.000000	1632.0
mean	0.202965	0.646359	0.179301	0.151202	0.140085	0.160131	-1.447462	4.301136	6.052402	0.0
std	0.177125	0.239934	0.163531	0.145799	0.120520	0.133439	1.064618	1.585021	1.913312	0.0
min	0.047619	0.047619	0.047619	0.047619	0.047619	0.047619	-4.295169	0.000000	1.666667	0.0
25%	0.047619	0.476190	0.047619	0.047619	0.047619	0.047619	-1.989108	3.208805	4.597059	0.0
50%	0.142857	0.666667	0.095238	0.095238	0.095238	0.095238	-1.287389	4.129448	5.750000	0.0
75%	0.285714	0.857143	0.238095	0.190476	0.190476	0.190476	-0.999840	5.255947	7.333333	0.0
max	1.000000	1.000000	0.904762	0.904762	0.761905	0.904762	0.987624	8.927174	11.181818	0.0

# get the second cluster's statistical result
df_kmeans[df_kmeans['KMEANS'] == 1][['F1' , 'F2' , 'O1' , 'D1' , 'O2' , 'D2' , 'FS' , 'STD_FD' , 'AVG_FD' , 'KMEANS']].describe()

	F1	F2	O1	D1	O2	D2	FS	STD_FD	AVG_FD	KMEANS
count	1812.000000	1812.000000	1812.000000	1812.000000	1812.000000	1812.000000	1812.000000	1812.000000	1812.000000	1812.0
mean	0.111768	0.393173	0.122753	0.116367	0.108615	0.116446	-0.949312	1.237290	2.780979	1.0
std	0.094105	0.235192	0.111258	0.109758	0.094724	0.099869	0.992366	1.143362	1.772891	0.0
min	0.047619	0.047619	0.047619	0.047619	0.047619	0.047619	-4.289637	0.000000	0.000000	1.0
25%	0.047619	0.190476	0.047619	0.047619	0.047619	0.047619	-1.454488	0.000000	1.567460	1.0
50%	0.047619	0.380952	0.095238	0.047619	0.095238	0.095238	-1.028056	1.154701	2.666667	1.0
75%	0.142857	0.571429	0.142857	0.142857	0.142857	0.142857	-0.301323	2.081666	3.839286	1.0
max	0.666667	1.000000	0.761905	0.857143	0.714286	0.619048	0.992855	5.656854	10.000000	1.0

# get the second cluster's statistical result
df_kmeans[df_kmeans['KMEANS'] == 2][['F1' , 'F2' , 'O1' , 'D1' , 'O2' , 'D2' , 'FS' , 'STD_FD' , 'AVG_FD' , 'KMEANS']].describe()

	F1	F2	O1	D1	O2	D2	FS	STD_FD	AVG_FD	KMEANS
count	1179.000000	1179.000000	1179.000000	1179.000000	1179.000000	1179.000000	1179.000000	1179.000000	1179.000000	1179.0
mean	0.606123	0.877620	0.573367	0.551799	0.492427	0.497354	-0.490052	2.142129	3.532619	2.0
std	0.210299	0.111866	0.201344	0.200316	0.196009	0.195912	0.995087	1.470733	1.747467	0.0
min	0.047619	0.476190	0.047619	0.047619	0.047619	0.047619	-4.109232	0.000000	0.000000	2.0
25%	0.476190	0.809524	0.428571	0.428571	0.380952	0.380952	-1.151578	1.100124	2.333333	2.0
50%	0.619048	0.904762	0.571429	0.571429	0.476190	0.476190	-0.526925	2.081666	3.400000	2.0
75%	0.761905	0.952381	0.714286	0.714286	0.619048	0.666667	0.384684	3.008290	4.666667	2.0
max	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000	0.996411	8.114141	10.000000	2.0

Start ISODATA Clustering based on the parameters that get from the kmeans

import isodata_cluster
importlib.reload(isodata_cluster)
# initial_set为初始设置集合，依次为nc,K,min_num,s,c,L,I,k
initial_set = [3 , 3 ,800 , 0.95 , 1.6 , 1 , 30 , 0.5]
isodata = isodata_cluster.IsoData(initial_set = initial_set, data = df_feat_new[['f1' , 'f2' , 'f3']].values)
result = isodata.result
# 将ISODATA的结果可视化展示出来
plt.rcParams['figure.dpi'] = 150 #分辨率
cluster_num = []   #记录各个簇的样本数量
cluster_sta = []   #记录各个簇的标准差最大分量
cluster_cen = []   #记录各个簇的簇中心点

fig = plt.figure()
ax = Axes3D(fig)
ax.set_xlabel("Factor 1") ; ax.set_ylabel("Factor 2") ; ax.set_zlabel("Factor 3") 

for cluster in result:
    df_clu = pd.DataFrame(cluster)
    df_clu.rename(columns = {0 : 'f1' , 1 : 'f2' , 2 : 'f3'} , inplace = True)
    cluster_num.append(df_clu.shape[0])
    cluster_sta.append(get_dis_sta(df_clu))
    cluster_cen.append([df_clu.mean()[0] , df_clu.mean()[1] , df_clu.mean()[2]])
    ax.scatter(df_clu['f1'] , df_clu['f2'] , df_clu['f3'] , s = 1, c=None )
    
dis_cen = get_cen_dis(cluster_cen)
print("各个簇的样本个数:" , cluster_num)
print("各个簇的标准差最大分量:" , cluster_sta)
#print("各个簇的中心坐标值:" , cluster_cen)
print("各个簇中心间的距离:" , dis_cen)
plt.show()

各个簇的样本个数: [1005, 2199, 1419]
各个簇的标准差最大分量: [0.9748475138394216, 0.850603792799397, 1.009029401333998]
各个簇中心间的距离: [1.7606495274478837, 1.9373602475202636, 1.9365002523817587]

是否需要修改参数,输入“y”确认修改,“n”取消
n
各个簇的样本个数: [1250, 2096, 1277]
各个簇的标准差最大分量: [0.9947203326867352, 0.843380563984188, 0.9317069892509775]
各个簇中心间的距离: [1.7709978211378359, 2.1006969090778655, 1.9934786038768135]

是否需要修改参数,输入“y”确认修改,“n”取消
n
各个簇的样本个数: [1390, 2008, 1225]
各个簇的标准差最大分量: [0.9649615589415986, 0.8227809146148534, 0.8784324812455603]
各个簇中心间的距离: [1.7586905358574996, 2.1303915270221165, 2.0447099723958706]

是否需要修改参数,输入“y”确认修改,“n”取消
n
各个簇的样本个数: [1474, 1951, 1198]
各个簇的标准差最大分量: [0.9349691503220108, 0.8005593708164274, 0.8519235775400618]
各个簇中心间的距离: [1.7521613158511906, 2.13940448447641, 2.0766470060066746]

是否需要修改参数,输入“y”确认修改,“n”取消
n
各个簇的样本个数: [1530, 1908, 1185]
各个簇的标准差最大分量: [0.912198664004, 0.8007782918360405, 0.8572159176093301]
各个簇中心间的距离: [1.7473163489833492, 2.1423171269973182, 2.096578306952844]

是否需要修改参数,输入“y”确认修改,“n”取消
n
各个簇的样本个数: [1564, 1876, 1183]
各个簇的标准差最大分量: [0.9098318586046931, 0.813456286538858, 0.8817334737847586]
各个簇中心间的距离: [1.745647496131135, 2.1413414911458655, 2.1050320412462344]

是否需要修改参数,输入“y”确认修改,“n”取消
n
各个簇的样本个数: [1591, 1851, 1181]
各个簇的标准差最大分量: [0.9194505945989653, 0.8262930052323401, 0.8962410538489356]
各个簇中心间的距离: [1.7424415048334665, 2.1412781020389136, 2.112605262873411]

是否需要修改参数,输入“y”确认修改,“n”取消
n
各个簇的样本个数: [1598, 1847, 1178]
各个簇的标准差最大分量: [0.9304903811761279, 0.8400875716377163, 0.9051774387075697]
各个簇中心间的距离: [1.744567685043581, 2.1382893793659425, 2.1181434099655148]

是否需要修改参数,输入“y”确认修改,“n”取消
n
各个簇的样本个数: [1638, 1111, 1180]
各个簇的标准差最大分量: [0.8279349006048432, 0.8831351172545858, 0.6941034796157513]
各个簇中心间的距离: [2.174220849826817, 1.7848643561500932, 2.1892024134710346]

是否需要修改参数,输入“y”确认修改,“n”取消
n
各个簇的样本个数: [1648, 1112, 896, 967]
各个簇的标准差最大分量: [0.8940802214936883, 0.93678246937646, 0.6854077115112495, 0.794068466133068]
各个簇中心间的距离: [2.1687733997851177, 1.9187235189509044, 1.9468007867484385, 2.2781365940311744, 2.266002811324093, 1.7070120289999497]

是否需要修改参数,输入“y”确认修改,“n”取消
n
各个簇的样本个数: [1534, 1051, 971, 1067]
各个簇的标准差最大分量: [0.8074138457404908, 0.8869132242334714, 0.7385445061857353, 0.8017438096912138]
各个簇中心间的距离: [2.21480762458012, 1.9098940572182783, 1.9145867586823786, 2.3050541551203025, 2.2503765665682733, 1.8336856589421984]

是否需要修改参数,输入“y”确认修改,“n”取消
n
各个簇的样本个数: [1477, 1015, 1014, 1117]
各个簇的标准差最大分量: [0.7770073579955322, 0.8641096734201485, 0.7598582883841911, 0.8016173055422908]
各个簇中心间的距离: [2.25044508096158, 1.9067469069258582, 1.8938868415271946, 2.310618906820702, 2.246246206909459, 1.8517643463373639]

是否需要修改参数,输入“y”确认修改,“n”取消
n
各个簇的样本个数: [1453, 996, 1062, 1112]
各个簇的标准差最大分量: [0.7705164251046231, 0.8517702999683567, 0.7646669741811046, 0.7980086209523115]
各个簇中心间的距离: [2.266735571956418, 1.9100670200444594, 1.8840899661058563, 2.3187067678599025, 2.2477941178118286, 1.8529037135290922]

是否需要修改参数,输入“y”确认修改,“n”取消
n
各个簇的样本个数: [1428, 993, 1114, 1088]
各个簇的标准差最大分量: [0.7661732280020518, 0.8388934251456486, 0.7642656587898279, 0.7937322088731378]
各个簇中心间的距离: [2.2805502917951355, 1.922449339545091, 1.8734666650342908, 2.3163151592358737, 2.244396872048285, 1.8631823963382783]

是否需要修改参数,输入“y”确认修改,“n”取消
n
各个簇的样本个数: [1407, 991, 1176, 1049]
各个簇的标准差最大分量: [0.7467885094340987, 0.8320938674584458, 0.7567616870044901, 0.7840233353587531]
各个簇中心间的距离: [2.281207775356509, 1.9213705440494986, 1.8823533796548302, 2.3109522927451547, 2.255181459580774, 1.8925832018514959]

是否需要修改参数,输入“y”确认修改,“n”取消
n
各个簇的样本个数: [1398, 998, 1223, 1004]
各个簇的标准差最大分量: [0.7370544467537546, 0.8287788247634427, 0.7381212913701396, 0.7733264193449961]
各个簇中心间的距离: [2.274357473838571, 1.9195485219912651, 1.8996994821051885, 2.299312432584461, 2.276691305879818, 1.9227540506311296]

是否需要修改参数,输入“y”确认修改,“n”取消
n
各个簇的样本个数: [1395, 1005, 1261, 962]
各个簇的标准差最大分量: [0.7312847767631867, 0.8258615601365257, 0.7176429804529262, 0.7831896886772651]
各个簇中心间的距离: [2.2704736799095815, 1.9118253005710208, 1.9250797144646383, 2.282421119399488, 2.3015048787128958, 1.954050272305322]

是否需要修改参数,输入“y”确认修改,“n”取消
n
各个簇的样本个数: [1404, 1007, 1280, 932]
各个簇的标准差最大分量: [0.7276366390182321, 0.821684999427722, 0.7122941386352414, 0.7938630439931957]
各个簇中心间的距离: [2.2671988264057217, 1.903220983238355, 1.947190196017278, 2.2717424670874493, 2.3195476386706106, 1.9799902758637797]

是否需要修改参数,输入“y”确认修改,“n”取消
n
各个簇的样本个数: [1412, 1003, 1289, 919]
各个簇的标准差最大分量: [0.7266708924018204, 0.8140573204739283, 0.7172455897384683, 0.7842865714538344]
各个簇中心间的距离: [2.2624126581552937, 1.9057340633747935, 1.9561905244584867, 2.27019481397084, 2.3326023463398613, 1.9938032878922927]

是否需要修改参数,输入“y”确认修改,“n”取消
n
各个簇的样本个数: [1423, 1008, 1296, 896]
各个簇的标准差最大分量: [0.7265455707569093, 0.815763113421513, 0.7216894162239375, 0.7803219667937547]
各个簇中心间的距离: [2.2565843345841756, 1.90727173450547, 1.964631875475134, 2.264586331828808, 2.3451101027635888, 2.0045236800208923]

是否需要修改参数,输入“y”确认修改,“n”取消
n
各个簇的样本个数: [1432, 1008, 1309, 874]
各个簇的标准差最大分量: [0.7258379535753273, 0.8163351430269637, 0.7268937532780968, 0.7752710044935526]
各个簇中心间的距离: [2.253428141589837, 1.9040523417118924, 1.9736009168747597, 2.263975558182213, 2.3535200915298806, 2.01740535531187]

是否需要修改参数,输入“y”确认修改,“n”取消
n
各个簇的样本个数: [1443, 1005, 1312, 863]
各个簇的标准差最大分量: [0.7257874258379888, 0.8159790017872471, 0.72788424443682, 0.7765188694672911]
各个簇中心间的距离: [2.253319283599768, 1.8989065925886106, 1.9791669184587413, 2.2656480889494484, 2.3599066497721677, 2.02712036819384]

是否需要修改参数,输入“y”确认修改,“n”取消
n
各个簇的样本个数: [1449, 1005, 1312, 857]
各个簇的标准差最大分量: [0.7249767767202866, 0.8149561628048259, 0.728351642655314, 0.7785733489018503]
各个簇中心间的距离: [2.2502042556862674, 1.8963675662703723, 1.983880302673049, 2.265978390859143, 2.3628690769452914, 2.0335023454878582]

是否需要修改参数,输入“y”确认修改,“n”取消
n
各个簇的样本个数: [1452, 1008, 1319, 844]
各个簇的标准差最大分量: [0.7248178092487785, 0.8143841810831114, 0.7349563295485158, 0.7691303611627518]
各个簇中心间的距离: [2.2484306735869555, 1.8964451535604525, 1.9854111633970135, 2.2663900323748205, 2.3614996332233407, 2.0422881482353734]

是否需要修改参数,输入“y”确认修改,“n”取消
n
各个簇的样本个数: [1452, 1013, 1322, 836]
各个簇的标准差最大分量: [0.724260303681589, 0.814006827698086, 0.7346917392677543, 0.7693270789730574]
各个簇中心间的距离: [2.243309680107007, 1.8948118314469045, 1.9918486966721136, 2.2632922644769033, 2.3653814145464307, 2.0475208270863763]

是否需要修改参数,输入“y”确认修改,“n”取消
n
各个簇的样本个数: [1455, 1012, 1329, 827]
各个簇的标准差最大分量: [0.7251678330407104, 0.8114617328007601, 0.7366361349161656, 0.7664510457579726]
各个簇中心间的距离: [2.2416205519544796, 1.8926066555889862, 1.9975094634993529, 2.261922722766794, 2.3689959541955274, 2.054641027981982]

是否需要修改参数,输入“y”确认修改,“n”取消
n
各个簇的样本个数: [1458, 1013, 1332, 820]
各个簇的标准差最大分量: [0.7251776660480466, 0.8116677350958016, 0.738610754885011, 0.7641389095449126]
各个簇中心间的距离: [2.2392018856006786, 1.8920736602221055, 2.000050857311144, 2.2622501448226044, 2.3702837221731525, 2.0597262908748326]

是否需要修改参数,输入“y”确认修改,“n”取消
n
各个簇的样本个数: [1458, 1012, 1335, 818]
各个簇的标准差最大分量: [0.7248220020374851, 0.8108737843044832, 0.7443489260010758, 0.7550721141513254]
各个簇中心间的距离: [2.2403664412418705, 1.8941438192089173, 1.9959611827221577, 2.2636172001615837, 2.3668572911917285, 2.0630510423342048]

各个簇的样本个数: [1458, 1012, 1335, 818]
各个簇的标准差最大分量: [0.7248220020374851, 0.8108737843044832, 0.7443489260010758, 0.7550721141513254]
各个簇中心间的距离: [2.2403664412418705, 1.8941438192089173, 1.9959611827221577, 2.2636172001615837, 2.3668572911917285, 2.0630510423342048]

# get the isodata's result's Silhouette Coefficient
df_isodata = pd.DataFrame(columns = ['f1' , 'f2' , 'f3' , 'ISODATA']) 
for i in range(len(result)):
    df_temp = pd.DataFrame(result[i])
    df_temp.rename(columns = {0 : 'f1' , 1 : 'f2' , 2 : 'f3'} , inplace = True)
    df_temp['ISODATA'] = i
    df_isodata = df_isodata.append(df_temp)
df_isodata.reset_index(drop = True , inplace = True)
# 实际上下面这种表联结的方式存在一些问题，因为如果有几辆车的factor1 , factor2 , factor3完全相同的话，
# 那么就会出现他们彼此间两两都会联结的情况了，这就导致样本集的数量增多了，但是这种情况是可以接受的，
# 因为第一出现这种情况的概率是很小的，第二如果两辆车的feature都相同，则必定会被聚到一个类，直接去重是不会影响的。
df_isodata = pd.merge(df_feat_new , df_isodata , on = ['f1' , 'f2' , 'f3'] , how = 'inner')
df_isodata.drop_duplicates(subset=['CAR_NUM'], keep='first', inplace = True)
df_isodata['ISODATA'] = df_isodata['ISODATA'].astype(np.int)  # 将ISODATA这一列转化为np.int类型，使得describe()可以成功对这一列做统计
df_isodata.head(5)

	f1	f2	f3	CAR_NUM	DBSCAN	F1	F2	O1	D1	O2	D2	FS	STD_FD	AVG_FD	FastSearch	ISODATA
0	0.633961	0.057235	-1.165552	浙AFZ391	1	0.047619	0.904762	0.142857	0.142857	0.571429	0.619048	-2.000000	2.093641	5.125000	1	3
1	-0.878013	-0.763536	-0.307468	浙B732UE	0	0.047619	0.428571	0.047619	0.047619	0.047619	0.047619	-1.453860	1.118034	2.333333	0	0
2	1.312621	0.156470	0.043351	浙AN0G37	-1	0.428571	0.952381	0.523810	0.523810	0.523810	0.523810	-0.677038	2.547726	4.909091	1	1
3	-0.335299	-0.023774	-0.200037	浙A3WA22	0	0.047619	0.571429	0.047619	0.047619	0.285714	0.333333	-1.199703	2.958040	3.666667	0	0
4	0.847105	1.263833	-0.014173	浙GN820X	-1	0.476190	0.857143	0.476190	0.285714	0.285714	0.523810	-0.820671	5.780846	5.727273	1	2

# get the Silhouette Coefficient of the isodata's result.
silhouettevalue = metrics.silhouette_score(df_isodata[['f1' , 'f2' , 'f3']].values , df_isodata['ISODATA'] , metric='euclidean')
silhouettevalue

0.2854265503769039

Understanding every clusters' statistical characteristic

# get the first cluster's statistical result
df_isodata[df_isodata['ISODATA'] == 0][['F1' , 'F2' , 'O1' , 'D1' , 'O2' , 'D2' , 'FS' , 'STD_FD' , 'AVG_FD' , 'ISODATA']].describe()

	F1	F2	O1	D1	O2	D2	FS	STD_FD	AVG_FD	ISODATA
count	1458.000000	1458.000000	1458.000000	1458.000000	1458.000000	1458.000000	1458.000000	1458.000000	1458.000000	1458.0
mean	0.111111	0.373702	0.119244	0.113855	0.108400	0.113103	-0.778688	1.109191	2.657246	0.0
std	0.095205	0.227441	0.106986	0.106393	0.094513	0.096760	0.771450	1.050375	1.722388	0.0
min	0.047619	0.047619	0.047619	0.047619	0.047619	0.047619	-2.539649	0.000000	0.000000	0.0
25%	0.047619	0.190476	0.047619	0.047619	0.047619	0.047619	-1.265825	0.000000	1.500000	0.0
50%	0.047619	0.333333	0.095238	0.047619	0.095238	0.095238	-1.006318	1.000000	2.594118	0.0
75%	0.142857	0.523810	0.142857	0.142857	0.142857	0.142857	-0.295296	1.966384	3.666667	0.0
max	0.666667	1.000000	0.761905	0.857143	0.714286	0.619048	0.992855	4.500000	9.000000	0.0

# get the second cluster's statistical result
df_isodata[df_isodata['ISODATA'] == 1][['F1' , 'F2' , 'O1' , 'D1' , 'O2' , 'D2' , 'FS' , 'STD_FD' , 'AVG_FD' , 'ISODATA']].describe()

	F1	F2	O1	D1	O2	D2	FS	STD_FD	AVG_FD	ISODATA
count	1012.000000	1012.000000	1012.000000	1012.000000	1012.000000	1012.000000	1012.000000	1012.000000	1012.000000	1012.0
mean	0.631470	0.882364	0.593356	0.575805	0.512187	0.517034	-0.374766	2.083573	3.470499	1.0
std	0.203464	0.109400	0.194631	0.193829	0.191233	0.191682	0.854781	1.461300	1.724281	0.0
min	0.047619	0.476190	0.047619	0.047619	0.047619	0.047619	-2.377553	0.000000	0.000000	1.0
25%	0.476190	0.809524	0.476190	0.428571	0.380952	0.380952	-1.049218	1.000000	2.326923	1.0
50%	0.666667	0.904762	0.619048	0.571429	0.523810	0.523810	-0.446112	2.049390	3.333333	1.0
75%	0.761905	1.000000	0.714286	0.714286	0.666667	0.666667	0.411075	2.968276	4.637987	1.0
max	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000	0.996411	8.082904	10.750000	1.0

# get the third cluster's statistical result
df_isodata[df_isodata['ISODATA'] == 2][['F1' , 'F2' , 'O1' , 'D1' , 'O2' , 'D2' , 'FS' , 'STD_FD' , 'AVG_FD' , 'ISODATA']].describe()

	F1	F2	O1	D1	O2	D2	FS	STD_FD	AVG_FD	ISODATA
count	1335.000000	1335.000000	1335.000000	1335.000000	1335.000000	1335.000000	1335.000000	1335.000000	1335.000000	1335.0
mean	0.222008	0.640663	0.187587	0.159979	0.143428	0.166970	-0.808962	4.529741	6.431054	2.0
std	0.189692	0.243896	0.168152	0.150288	0.120821	0.131207	0.809043	1.539201	1.813513	0.0
min	0.047619	0.047619	0.047619	0.047619	0.047619	0.047619	-3.798238	0.000000	2.875000	2.0
25%	0.047619	0.476190	0.047619	0.047619	0.047619	0.095238	-1.307800	3.430258	5.000000	2.0
50%	0.142857	0.666667	0.142857	0.095238	0.095238	0.095238	-1.008589	4.356850	6.142857	2.0
75%	0.333333	0.857143	0.285714	0.190476	0.190476	0.238095	-0.256081	5.501623	7.600000	2.0
max	1.000000	1.000000	1.000000	0.904762	0.809524	0.761905	0.992855	8.927174	11.181818	2.0

# get the third cluster's statistical result
df_isodata[df_isodata['ISODATA'] == 3][['F1' , 'F2' , 'O1' , 'D1' , 'O2' , 'D2' , 'FS' , 'STD_FD' , 'AVG_FD' , 'ISODATA']].describe()

	F1	F2	O1	D1	O2	D2	FS	STD_FD	AVG_FD	ISODATA
count	818.000000	818.000000	818.000000	818.000000	818.000000	818.000000	818.000000	818.000000	818.000000	818.0
mean	0.184538	0.622133	0.203283	0.178368	0.168879	0.180522	-2.525216	2.462119	3.801645	3.0
std	0.171614	0.261615	0.193704	0.176193	0.162676	0.169826	0.812167	1.534517	1.761414	0.0
min	0.047619	0.047619	0.047619	0.047619	0.047619	0.047619	-4.295169	0.000000	0.000000	3.0
25%	0.047619	0.428571	0.047619	0.047619	0.047619	0.047619	-3.114484	1.486436	2.666667	3.0
50%	0.095238	0.666667	0.095238	0.095238	0.095238	0.095238	-2.327513	2.449490	3.700000	3.0
75%	0.238095	0.809524	0.333333	0.238095	0.238095	0.238095	-1.913535	3.329570	4.794444	3.0
max	0.952381	1.000000	0.857143	0.809524	0.952381	0.904762	-1.026099	8.897771	10.666667	3.0

Clustering by fast search and find of density peaks

Decision Graph Plot

Get the decision graph that is proposed by the paper "Clustering by fast search and find of density peaks". The graph can determine the optimal clustering number. Here i will not determine the optimal clustering number based on this method, but i also plot the decision graph

import ClusteringByFastSearchAndFindOfDensityPeaks
importlib.reload(ClusteringByFastSearchAndFindOfDensityPeaks)

data = df_feat_new[['f1' , 'f2' , 'f3']].values
N = len(data)
dist_sample , dc , sample_density , density_rank , n_list , delta_sample = \
    ClusteringByFastSearchAndFindOfDensityPeaks.step1(data , 0.01)
plt.rcParams['font.sans-serif'] = ['SimHei']  #显示中文字符
plt.rcParams['axes.unicode_minus']=False  # 用来正常显示负号
plt.scatter(sample_density , delta_sample , s = 10 , marker = '.')
plt.grid(True)
plt.xlabel('Density')
plt.ylabel('Distance')
plt.title("Decision Graph")
plt.show()

#识别聚类数目
density_delta_multi = np.multiply(np.array(sample_density),np.array(delta_sample))
density_delta_multi = density_delta_multi.tolist()  #样本的密度数组与样本的距离数组中的对应元素相乘,生成density_delta_multi数组
density_delta_multi_copy = copy.deepcopy(density_delta_multi)  #对density_delta_multi数组进行深拷贝
density_delta_multi_copy.sort(reverse = True)  #对系数按照降序排列
x_axis = [ i for i in range(N)]
plt.scatter(x_axis , density_delta_multi_copy , s = 10 , marker = '.')
plt.grid(True)
plt.xlabel('The order of point')
plt.ylabel('Coefficient')
plt.title("聚类数目识别图")
plt.show()

Start clustering by fast search and find of density peaks, and display the result

# 进行聚类操作
cluster_center_index , tag_list , tag_core_halo = \
    ClusteringByFastSearchAndFindOfDensityPeaks.fast_search_and_find_density_peaks(\
        data , cluster_num = 3 , t_percent = 0.015)
tag_list = np.array(tag_list)
df_feat_new['FastSearch'] = tag_list

# 绘制出样本的聚类结果中的各个簇
plt.rcParams['font.sans-serif'] = ['SimHei']  #显示中文字符
plt.rcParams['axes.unicode_minus']=False  # 用来正常显示负号
# plt.rcParams['savefig.dpi'] = 300 #图片像素
plt.rcParams['figure.dpi'] = 150 #分辨率
fig = plt.figure()
ax = Axes3D(fig)
ax.set_xlabel('Factor 1')
ax.set_ylabel('Factor 2')
ax.set_zlabel('Factor 3')
for i in range(4):
    one_cluster = df_feat_new[['f1' , 'f2' , 'f3']].values[tag_list == i]
    ax.scatter(one_cluster[:,0],one_cluster[:,1],one_cluster[:,2] , s = 1 , c = None)
# plt.savefig("不包含噪声的样本数据分簇情况.png")
plt.show()

dc: 0.47402132145368003

Get every clusters' statistical infomation

# the first cluster's infomation
df_feat_new[['F1' , 'F2' , 'O1' , 'D1' , 'O2' , 'D2' , 'FS' , 'STD_FD' , 'AVG_FD' , 'FastSearch']][df_feat_new['FastSearch'] == 0].describe()

	F1	F2	O1	D1	O2	D2	FS	STD_FD	AVG_FD	FastSearch
count	2190.000000	2190.000000	2190.000000	2190.000000	2190.000000	2190.000000	2190.000000	2190.000000	2190.000000	2190.0
mean	0.203914	0.496434	0.213525	0.199674	0.187258	0.193107	-1.293739	1.266167	2.641789	0.0
std	0.230658	0.292899	0.230369	0.220156	0.204025	0.204194	1.059593	1.109374	1.545560	0.0
min	0.047619	0.047619	0.047619	0.047619	0.047619	0.047619	-4.295169	0.000000	0.000000	0.0
25%	0.047619	0.238095	0.047619	0.047619	0.047619	0.047619	-1.773418	0.000000	1.625000	0.0
50%	0.095238	0.476190	0.095238	0.095238	0.095238	0.095238	-1.168499	1.258306	2.636364	0.0
75%	0.238095	0.761905	0.285714	0.285714	0.238095	0.238095	-0.794355	2.081666	3.600000	0.0
max	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000	0.992855	5.737305	9.000000	0.0

# the second cluster's infomation
df_feat_new[['F1' , 'F2' , 'O1' , 'D1' , 'O2' , 'D2' , 'FS' , 'STD_FD' , 'AVG_FD' , 'FastSearch']][df_feat_new['FastSearch'] == 1].describe()

	F1	F2	O1	D1	O2	D2	FS	STD_FD	AVG_FD	FastSearch
count	2115.000000	2115.000000	2115.000000	2115.000000	2115.000000	2115.000000	2115.000000	2115.000000	2115.000000	2115.0
mean	0.268535	0.673804	0.242193	0.217629	0.195092	0.211753	-0.937688	3.944711	5.738507	1.0
std	0.221040	0.242915	0.209230	0.198685	0.170064	0.171627	0.979741	1.709317	2.000988	0.0
min	0.047619	0.047619	0.047619	0.047619	0.047619	0.047619	-4.272598	0.000000	0.000000	1.0
25%	0.095238	0.523810	0.095238	0.047619	0.047619	0.095238	-1.482893	2.878492	4.379808	1.0
50%	0.190476	0.714286	0.142857	0.142857	0.142857	0.142857	-1.024084	3.800585	5.437500	1.0
75%	0.428571	0.857143	0.380952	0.333333	0.285714	0.333333	-0.226371	4.961595	7.000000	1.0
max	1.000000	1.000000	1.000000	0.952381	0.904762	0.952381	0.996411	8.927174	11.181818	1.0

# the third cluster's infomation
df_feat_new[['F1' , 'F2' , 'O1' , 'D1' , 'O2' , 'D2' , 'FS' , 'STD_FD' , 'AVG_FD' , 'FastSearch']][df_feat_new['FastSearch'] == 2].describe()

	F1	F2	O1	D1	O2	D2	FS	STD_FD	AVG_FD	FastSearch
count	318.000000	318.000000	318.000000	318.000000	318.000000	318.000000	318.000000	318.000000	318.000000	318.0
mean	0.735400	0.911051	0.664121	0.662324	0.576370	0.591045	0.491565	2.110143	3.645135	2.0
std	0.148732	0.086767	0.161586	0.159408	0.171246	0.170775	0.376711	1.464915	1.702781	0.0
min	0.285714	0.619048	0.190476	0.238095	0.142857	0.095238	-0.593192	0.000000	0.000000	2.0
25%	0.619048	0.857143	0.571429	0.571429	0.476190	0.476190	0.197497	1.154701	2.616667	2.0
50%	0.761905	0.952381	0.666667	0.666667	0.571429	0.619048	0.562012	2.081666	3.563492	2.0
75%	0.857143	1.000000	0.761905	0.761905	0.714286	0.714286	0.833020	3.020530	4.770833	2.0
max	1.000000	1.000000	1.000000	1.000000	1.000000	0.952381	0.988726	6.626067	8.500000	2.0

Get the clusters' Silhouette Coefficient

# get the Silhouette Coefficient of the cluster's result.
silhouettevalue = metrics.silhouette_score(df_feat_new[['f1' , 'f2' , 'f3']].values , df_feat_new['FastSearch'] , metric='euclidean')
silhouettevalue

0.175708763238439

Get the commuting pattern rule based on decision tree model

train the decision model

df_isodata.head(5)

	f1	f2	f3	CAR_NUM	DBSCAN	F1	F2	O1	D1	O2	D2	FS	STD_FD	AVG_FD	FastSearch	ISODATA
0	0.633961	0.057235	-1.165552	浙AFZ391	1	0.047619	0.904762	0.142857	0.142857	0.571429	0.619048	-2.000000	2.093641	5.125000	1	3
1	-0.878013	-0.763536	-0.307468	浙B732UE	0	0.047619	0.428571	0.047619	0.047619	0.047619	0.047619	-1.453860	1.118034	2.333333	0	0
2	1.312621	0.156470	0.043351	浙AN0G37	-1	0.428571	0.952381	0.523810	0.523810	0.523810	0.523810	-0.677038	2.547726	4.909091	1	1
3	-0.335299	-0.023774	-0.200037	浙A3WA22	0	0.047619	0.571429	0.047619	0.047619	0.285714	0.333333	-1.199703	2.958040	3.666667	0	0
4	0.847105	1.263833	-0.014173	浙GN820X	-1	0.476190	0.857143	0.476190	0.285714	0.285714	0.523810	-0.820671	5.780846	5.727273	1	2

# df_isodata.to_csv("df_isodata.csv" , encoding = 'gbk')

I train the decision tree model based on the isodata result. And i get the best initial parameters of the decision tree model through several attempts, and finally i choose the initial parameters of " max_depth = 6 , min_samples_split = 50 , min_samples_leaf = 20".

df_isodata = pd.read_csv("df_isodata.csv" , encoding = 'gbk')
df_isodata.drop(columns=['Unnamed: 0'] , inplace = True)
df_isodata.head(5)

	f1	f2	f3	CAR_NUM	DBSCAN	F1	F2	O1	D1	O2	D2	FS	STD_FD	AVG_FD	FastSearch	ISODATA
0	0.633961	0.057235	-1.165552	浙AFZ391	1	0.047619	0.904762	0.142857	0.142857	0.571429	0.619048	-2.000000	2.093641	5.125000	1	3
1	-0.878013	-0.763536	-0.307468	浙B732UE	0	0.047619	0.428571	0.047619	0.047619	0.047619	0.047619	-1.453860	1.118034	2.333333	0	0
2	1.312621	0.156470	0.043351	浙AN0G37	-1	0.428571	0.952381	0.523810	0.523810	0.523810	0.523810	-0.677038	2.547726	4.909091	1	1
3	-0.335299	-0.023774	-0.200037	浙A3WA22	0	0.047619	0.571429	0.047619	0.047619	0.285714	0.333333	-1.199703	2.958040	3.666667	0	0
4	0.847105	1.263833	-0.014173	浙GN820X	-1	0.476190	0.857143	0.476190	0.285714	0.285714	0.523810	-0.820671	5.780846	5.727273	1	2

accurancy_list = []
precision_list = []
recall_list = []

# 绘制决策树模型的图
def plot_decision_tree(dot_data , filename):
    #该函数将生成的决策树可视化出来
    graph = pydotplus.graph_from_dot_data(dot_data)
    graph.get_nodes()[7].set_fillcolor("#FFF2DD")  #指定颜色
    Image(graph.create_png()) #画出这个树
    graph.write_png(filename+ ".png") #保存为png格式

# 计算模型的正确率
def get_correct_ratio(data_test_target , result_prediction):
    '''
    data_test_target:测试集的标签list
    result_prediction:利用建好的树对测试集进行测试后的结果
    return:correct_ratio，模型的正确率
    '''
    N = len(result_prediction)  #测试集样本总个数
    count = 0
    for i in range(len(data_test_target)):
        #print( result_prediction[i])
        if data_test_target[i] == result_prediction[i]:
            count = count + 1
    correct_ratio = count / N
    return correct_ratio

# 划分训练集,测试集
N = len(df_isodata) # total number of sample
train_num = round(0.8 * N)
test_num = N - train_num
df_train = df_isodata[0 : train_num]
df_test = df_isodata[train_num : N]
data_train_target = df_train['ISODATA'].tolist()
data_train_target = [int(i) for i in data_train_target]
for i in range(len(data_train_target)):
    if data_train_target[i] == 1:
        data_train_target[i] = 1
    else:
        data_train_target[i] = 0
data_train = df_train[['F1' , 'F2' , 'O1' , 'D1' , 'O2' , 'D2' , 'FS' , 'STD_FD' , 'AVG_FD']].values
#data_train = data_train.tolist()

data_test_target = df_test['ISODATA'].tolist()
data_test_target = [int(i) for i in data_test_target]
for i in range(len(data_test_target)):
    if data_test_target[i] == 1:
        data_test_target[i] = 1
    else:
        data_test_target[i] = 0
data_test = df_test[['F1' , 'F2' , 'O1' , 'D1' , 'O2' , 'D2' , 'FS' , 'STD_FD' , 'AVG_FD']].values
#data_test = data_test.tolist()

#网格搜索法加交叉验证获取最优决策树模型参数
params_grid = {'max_depth':[2,3,4,5,6,7,8,9,10],'min_samples_leaf':[2,5,10,20,40] , 'min_samples_split':[20,50,80,100],\
               'class_weight' : ['balanced'] , 'random_state':[0] }
estimator = tree.DecisionTreeClassifier()
grid_search = GridSearchCV(estimator,  param_grid =params_grid, n_jobs = -1, verbose = 2 ,cv=5)  
grid_search.fit(data_train , data_train_target)
print(grid_search.best_score_)
print(grid_search.best_params_)
print(grid_search.best_estimator_)

Fitting 5 folds for each of 180 candidates, totalling 900 fits

[Parallel(n_jobs=-1)]: Using backend LokyBackend with 8 concurrent workers.
[Parallel(n_jobs=-1)]: Done  34 tasks      | elapsed:    0.2s
[Parallel(n_jobs=-1)]: Done 768 tasks      | elapsed:    2.4s

0.9451040485681892
{'class_weight': 'balanced', 'max_depth': 9, 'min_samples_leaf': 2, 'min_samples_split': 20, 'random_state': 0}
DecisionTreeClassifier(class_weight='balanced', max_depth=9, min_samples_leaf=2,
                       min_samples_split=20, random_state=0)

[Parallel(n_jobs=-1)]: Done 900 out of 900 | elapsed:    2.7s finished

'''
该函数用来生成剪之后的决策树
data_train:训练集list
data_train_target:训练集标签
data_test:测试集
data_test_target:测试集标签
min_samples_split_num:表示分裂一个内部节点需要的做少样本数
min_samples_leaf_num:指定每个叶子节点需要的最少样本数
没有返回值,会生成一张决策树的png图片
'''
clf = tree.DecisionTreeClassifier(class_weight='balanced', max_depth=9, min_samples_leaf=2,
                       min_samples_split=20, random_state=0)
clf = clf.fit(data_train , data_train_target)
result_prediction = clf.predict(data_test)
correct_ratio = get_correct_ratio(data_test_target , result_prediction)
print("The correct ratio of the mpdel is : " , correct_ratio)
car_feature_names = ['F1' , 'F2' , 'O1' , 'D1' , 'O2' , 'D2' , 'FS' , 'STD_FD' , 'AVG_FD' ]
car_target_names = ['0' , '1' ]
dot_data = tree.export_graphviz(clf, out_file=None, 
                      feature_names = car_feature_names,  
                      class_names=car_target_names,  
                      filled=True, rounded=True,  
                      special_characters=True)  
plot_decision_tree(dot_data , '决策树结果示意图')
# 保存模型
joblib.dump(clf,  "DecisionTree.pkl")   
# lr是训练好的模型， "./ML/test.pkl"是模型要保存的路径及保存模型的文件名，其中，'pkl' 是sklearn中默认的保存格式gai

The correct ratio of the mpdel is :  0.9491891891891892

['DecisionTree.pkl']

# 得到准确率
accurancy_list.append(accuracy_score(data_test_target, result_prediction))
precision_list.append(precision_score(data_test_target, result_prediction, average='macro'))
recall_list.append(recall_score(data_test_target, result_prediction, average='macro'))
print('accurancy:' ,accuracy_score(data_test_target, result_prediction, normalize=True))
print('macre-precision:' ,precision_score(data_test_target, result_prediction, average='macro')) # precision
print('macro-recall:' , recall_score(data_test_target,result_prediction, average='macro'))  # recall

accurancy: 0.9491891891891892
macre-precision: 0.9151174440610041
macro-recall: 0.9485445457218464

特征重要性排序

plt.rcParams['figure.dpi'] = 300
importance = clf.feature_importances_
# Impt_Series = pd.Series(importance, index = df_train[['F1' , 'F2' , 'O1' , 'D1' , 'O2' , 'D2' , 'FS' , 'STD_FD' , 'AVG_FD']].columns)
Impt_Series = pd.Series(importance, index = ['$F_1$' ,'$F_2$' , '$F_{O1}$' , '$F_{D1}$' , '$F_{O2}$' , '$F_{D2}$' , '$F_s$' , '$\sigma$' , '$M$'])
Impt_Series.sort_values(ascending = True).plot(kind='barh')
plt.xlabel("Feature Importance")
plt.show()

不同数目特征下的决策树准确率变化情况

# 8个特征训练决策树模型
data_train_tmp = data_train[: , [0,2,3,4,5,6,7,8]]
data_test_tmp = data_test[: , [0,2,3,4,5,6,7,8]]
#网格搜索法加交叉验证获取最优决策树模型参数
params_grid = {'max_depth':[2,3,4,5,6,7,8,9,10],'min_samples_leaf':[2,5,10,20,40] , 'min_samples_split':[20,50,80,100],\
               'class_weight' : ['balanced'] , 'random_state':[0] }
estimator = tree.DecisionTreeClassifier()
grid_search = GridSearchCV(estimator,  param_grid =params_grid, n_jobs = -1, verbose = 2 ,cv=5)  
grid_search.fit(data_train_tmp , data_train_target)
print(grid_search.best_score_)
print(grid_search.best_params_)
print(grid_search.best_estimator_)

Fitting 5 folds for each of 180 candidates, totalling 900 fits

[Parallel(n_jobs=-1)]: Using backend LokyBackend with 8 concurrent workers.
[Parallel(n_jobs=-1)]: Done  34 tasks      | elapsed:    0.2s

0.9456416633141937
{'class_weight': 'balanced', 'max_depth': 9, 'min_samples_leaf': 2, 'min_samples_split': 20, 'random_state': 0}
DecisionTreeClassifier(class_weight='balanced', max_depth=9, min_samples_leaf=2,
                       min_samples_split=20, random_state=0)

[Parallel(n_jobs=-1)]: Done 861 tasks      | elapsed:    2.5s
[Parallel(n_jobs=-1)]: Done 900 out of 900 | elapsed:    2.5s finished

clf8 = tree.DecisionTreeClassifier(class_weight='balanced', max_depth=9, min_samples_leaf=2,
                       min_samples_split=20, random_state=0)
clf8 = clf8.fit(data_train_tmp , data_train_target)
result_prediction = clf8.predict(data_test_tmp)
correct_ratio = get_correct_ratio(data_test_target , result_prediction)
print("The correct ratio of the mpdel is : " , correct_ratio)
car_feature_names = ['F1' , 'O1' , 'D1' , 'O2' , 'D2' ,'FS', 'STD_FD','AVG_FD']
car_target_names = ['0' , '1' ]
dot_data = tree.export_graphviz(clf8, out_file=None, 
                      feature_names = car_feature_names,  
                      class_names=car_target_names,  
                      filled=True, rounded=True,  
                      special_characters=True)  
plot_decision_tree(dot_data , '决策树结果示意图8')
# 保存模型
joblib.dump(clf8,  "DecisionTree8.pkl")   
# lr是训练好的模型， "./ML/test.pkl"是模型要保存的路径及保存模型的文件名，其中，'pkl' 是sklearn中默认的保存格式gai

The correct ratio of the mpdel is :  0.9535135135135135

['DecisionTree8.pkl']

# 得到准确率
accurancy_list.append(accuracy_score(data_test_target, result_prediction))
precision_list.append(precision_score(data_test_target, result_prediction, average='macro'))
recall_list.append(recall_score(data_test_target, result_prediction, average='macro'))
print('accurancy:' ,accuracy_score(data_test_target, result_prediction, normalize=True))
print('macre-precision:' ,precision_score(data_test_target, result_prediction, average='macro')) # precision
print('macro-recall:' , recall_score(data_test_target,result_prediction, average='macro'))  # recall

accurancy: 0.9535135135135135
macre-precision: 0.9208381263374028
macro-recall: 0.9547258827617546

# 7个特征训练决策树模型
data_train_tmp = data_train[: , [2,3,4,5,6,7,8]]
data_test_tmp = data_test[: ,  [2,3,4,5,6,7,8]]
#网格搜索法加交叉验证获取最优决策树模型参数
params_grid = {'max_depth':[2,3,4,5,6,7,8,9,10],'min_samples_leaf':[2,5,10,20,40] , 'min_samples_split':[20,50,80,100],\
               'class_weight' : ['balanced']  , 'random_state':[0]}
estimator = tree.DecisionTreeClassifier()
grid_search = GridSearchCV(estimator,  param_grid =params_grid, n_jobs = -1, verbose = 2 ,cv=5)  
grid_search.fit(data_train_tmp , data_train_target)
print(grid_search.best_score_)
print(grid_search.best_params_)
print(grid_search.best_estimator_)

Fitting 5 folds for each of 180 candidates, totalling 900 fits

[Parallel(n_jobs=-1)]: Using backend LokyBackend with 8 concurrent workers.
[Parallel(n_jobs=-1)]: Done  34 tasks      | elapsed:    0.3s
[Parallel(n_jobs=-1)]: Done 793 tasks      | elapsed:    2.5s

0.9405032366602055
{'class_weight': 'balanced', 'max_depth': 9, 'min_samples_leaf': 5, 'min_samples_split': 20, 'random_state': 0}
DecisionTreeClassifier(class_weight='balanced', max_depth=9, min_samples_leaf=5,
                       min_samples_split=20, random_state=0)

[Parallel(n_jobs=-1)]: Done 900 out of 900 | elapsed:    2.9s finished

clf7 = tree.DecisionTreeClassifier(class_weight='balanced', max_depth=9, min_samples_leaf=5,
                       min_samples_split=20, random_state=0)
clf7 = clf7.fit(data_train_tmp , data_train_target)
result_prediction = clf7.predict(data_test_tmp)
correct_ratio = get_correct_ratio(data_test_target , result_prediction)
print("The correct ratio of the mpdel is : " , correct_ratio)
car_feature_names = [ 'O1' , 'D1' , 'O2' , 'D2' ,'FS', 'STD_FD','AVG_FD']
car_target_names = ['0' , '1' ]
dot_data = tree.export_graphviz(clf7, out_file=None, 
                      feature_names = car_feature_names,  
                      class_names=car_target_names,  
                      filled=True, rounded=True,  
                      special_characters=True)  
plot_decision_tree(dot_data , '决策树结果示意图7')
# 保存模型
joblib.dump(clf7,  "DecisionTree7.pkl")   
# lr是训练好的模型， "./ML/test.pkl"是模型要保存的路径及保存模型的文件名，其中，'pkl' 是sklearn中默认的保存格式gai

The correct ratio of the mpdel is :  0.9427027027027027

['DecisionTree7.pkl']

# 得到准确率
accurancy_list.append(accuracy_score(data_test_target, result_prediction))
precision_list.append(precision_score(data_test_target, result_prediction, average='macro'))
recall_list.append(recall_score(data_test_target, result_prediction, average='macro'))
print('accurancy:' ,accuracy_score(data_test_target, result_prediction, normalize=True))
print('macre-precision:' ,precision_score(data_test_target, result_prediction, average='macro')) # precision
print('macro-recall:' , recall_score(data_test_target,result_prediction, average='macro'))  # recall

accurancy: 0.9427027027027027
macre-precision: 0.9044708029197079
macro-recall: 0.9443546015877683

# 6个特征训练决策树模型
data_train_tmp = data_train[: , [2,3,4,5,6,8]]
data_test_tmp = data_test[: , [2,3,4,5,6,8]]
#网格搜索法加交叉验证获取最优决策树模型参数
params_grid = {'max_depth':[2,3,4,5,6,7,8,9,10],'min_samples_leaf':[2,5,10,20,40] , 'min_samples_split':[20,50,80,100],\
               'class_weight' : ['balanced'] , 'random_state':[0] }
estimator = tree.DecisionTreeClassifier()
grid_search = GridSearchCV(estimator,  param_grid =params_grid, n_jobs = -1, verbose = 2 ,cv=5)  
grid_search.fit(data_train_tmp , data_train_target)
print(grid_search.best_score_)
print(grid_search.best_params_)
print(grid_search.best_estimator_)

Fitting 5 folds for each of 180 candidates, totalling 900 fits

[Parallel(n_jobs=-1)]: Using backend LokyBackend with 8 concurrent workers.
[Parallel(n_jobs=-1)]: Done  34 tasks      | elapsed:    0.2s
[Parallel(n_jobs=-1)]: Done 800 tasks      | elapsed:    2.4s

0.9407753355520608
{'class_weight': 'balanced', 'max_depth': 6, 'min_samples_leaf': 10, 'min_samples_split': 20, 'random_state': 0}
DecisionTreeClassifier(class_weight='balanced', max_depth=6,
                       min_samples_leaf=10, min_samples_split=20,
                       random_state=0)

[Parallel(n_jobs=-1)]: Done 900 out of 900 | elapsed:    2.7s finished

clf6 = tree.DecisionTreeClassifier(class_weight='balanced', max_depth=6,
                       min_samples_leaf=10, min_samples_split=20,
                       random_state=0)
clf6 = clf6.fit(data_train_tmp , data_train_target)
result_prediction = clf6.predict(data_test_tmp)
correct_ratio = get_correct_ratio(data_test_target , result_prediction)
print("The correct ratio of the mpdel is : " , correct_ratio)
car_feature_names = [ 'O1' , 'D1' , 'O2' , 'D2' ,'FS', 'AVG_FD']
car_target_names = ['0' , '1']
dot_data = tree.export_graphviz(clf6, out_file=None, 
                      feature_names = car_feature_names,  
                      class_names=car_target_names,  
                      filled=True, rounded=True,  
                      special_characters=True)  
plot_decision_tree(dot_data , '决策树结果示意图6')
# 保存模型
joblib.dump(clf6,  "DecisionTree6.pkl")   
# lr是训练好的模型， "./ML/test.pkl"是模型要保存的路径及保存模型的文件名，其中，'pkl' 是sklearn中默认的保存格式gai

The correct ratio of the mpdel is :  0.9437837837837838

['DecisionTree6.pkl']

# 得到准确率
accurancy_list.append(accuracy_score(data_test_target, result_prediction))
precision_list.append(precision_score(data_test_target, result_prediction, average='macro'))
recall_list.append(recall_score(data_test_target, result_prediction, average='macro'))
print('accurancy:' ,accuracy_score(data_test_target, result_prediction, normalize=True))
print('macre-precision:' ,precision_score(data_test_target, result_prediction, average='macro')) # precision
print('macro-recall:' , recall_score(data_test_target,result_prediction, average='macro'))  # recall

accurancy: 0.9437837837837838
macre-precision: 0.9043217286914766
macro-recall: 0.9501349870358986

# 5个特征训练决策树模型
data_train_tmp = data_train[: , [2,3,4,5,6]]
data_test_tmp = data_test[: , [2,3,4,5,6]]
#网格搜索法加交叉验证获取最优决策树模型参数
params_grid = {'max_depth':[2,3,4,5,6,7,8,9,10],'min_samples_leaf':[2,5,10,20,40] , 'min_samples_split':[20,50,80,100],\
               'class_weight' : ['balanced'] , 'random_state':[0]  }
estimator = tree.DecisionTreeClassifier()
grid_search = GridSearchCV(estimator,  param_grid =params_grid, n_jobs = -1, verbose = 2 ,cv=5)  
grid_search.fit(data_train_tmp , data_train_target)
print(grid_search.best_score_)
print(grid_search.best_params_)
print(grid_search.best_estimator_)

Fitting 5 folds for each of 180 candidates, totalling 900 fits

[Parallel(n_jobs=-1)]: Using backend LokyBackend with 8 concurrent workers.
[Parallel(n_jobs=-1)]: Done  34 tasks      | elapsed:    0.2s
[Parallel(n_jobs=-1)]: Done 800 tasks      | elapsed:    2.2s

0.9364506455034196
{'class_weight': 'balanced', 'max_depth': 6, 'min_samples_leaf': 5, 'min_samples_split': 20, 'random_state': 0}
DecisionTreeClassifier(class_weight='balanced', max_depth=6, min_samples_leaf=5,
                       min_samples_split=20, random_state=0)

[Parallel(n_jobs=-1)]: Done 900 out of 900 | elapsed:    2.6s finished

clf5 = tree.DecisionTreeClassifier(class_weight='balanced', max_depth=6, min_samples_leaf=5,
                       min_samples_split=20, random_state=0)
clf5 = clf5.fit(data_train_tmp , data_train_target)
result_prediction = clf5.predict(data_test_tmp)
correct_ratio = get_correct_ratio(data_test_target , result_prediction)
print("The correct ratio of the mpdel is : " , correct_ratio)
car_feature_names = [ 'O1' , 'D1' , 'O2' , 'D2' ,'FS']
car_target_names = ['0' , '1' ]
dot_data = tree.export_graphviz(clf5, out_file=None, 
                      feature_names = car_feature_names,  
                      class_names=car_target_names,  
                      filled=True, rounded=True,  
                      special_characters=True)  
plot_decision_tree(dot_data , '决策树结果示意图5')
# 保存模型
joblib.dump(clf5,  "DecisionTree5.pkl")   
# lr是训练好的模型， "./ML/test.pkl"是模型要保存的路径及保存模型的文件名，其中，'pkl' 是sklearn中默认的保存格式gai

The correct ratio of the mpdel is :  0.9459459459459459

['DecisionTree5.pkl']

# 得到准确率
accurancy_list.append(accuracy_score(data_test_target, result_prediction))
precision_list.append(precision_score(data_test_target, result_prediction, average='macro'))
recall_list.append(recall_score(data_test_target, result_prediction, average='macro'))
print('accurancy:' ,accuracy_score(data_test_target, result_prediction, normalize=True))
print('macre-precision:' ,precision_score(data_test_target, result_prediction, average='macro')) # precision
print('macro-recall:' , recall_score(data_test_target,result_prediction, average='macro'))  # recall

accurancy: 0.9459459459459459
macre-precision: 0.9104532839962998
macro-recall: 0.9447555531795462

# 4个特征训练决策树模型
data_train_tmp = data_train[: ,[2,3,4,5]]
data_test_tmp = data_test[: , [2,3,4,5]]
#网格搜索法加交叉验证获取最优决策树模型参数
params_grid = {'max_depth':[2,3,4,5,6,7,8,9,10],'min_samples_leaf':[2,5,10,20,40] , 'min_samples_split':[20,50,80,100],\
               'class_weight' : ['balanced'] , 'random_state':[0]  }
estimator = tree.DecisionTreeClassifier()
grid_search = GridSearchCV(estimator,  param_grid =params_grid, n_jobs = -1, verbose = 2 ,cv=5)  
grid_search.fit(data_train_tmp , data_train_target)
print(grid_search.best_score_)
print(grid_search.best_params_)
print(grid_search.best_estimator_)

Fitting 5 folds for each of 180 candidates, totalling 900 fits

[Parallel(n_jobs=-1)]: Using backend LokyBackend with 8 concurrent workers.
[Parallel(n_jobs=-1)]: Done  34 tasks      | elapsed:    0.2s

0.9250970998061663
{'class_weight': 'balanced', 'max_depth': 8, 'min_samples_leaf': 10, 'min_samples_split': 20, 'random_state': 0}
DecisionTreeClassifier(class_weight='balanced', max_depth=8,
                       min_samples_leaf=10, min_samples_split=20,
                       random_state=0)

[Parallel(n_jobs=-1)]: Done 832 tasks      | elapsed:    2.1s
[Parallel(n_jobs=-1)]: Done 900 out of 900 | elapsed:    2.2s finished

clf4 = tree.DecisionTreeClassifier(class_weight='balanced', max_depth=8,
                       min_samples_leaf=10, min_samples_split=20,
                       random_state=0)
clf4 = clf4.fit(data_train_tmp , data_train_target)
result_prediction = clf4.predict(data_test_tmp)
correct_ratio = get_correct_ratio(data_test_target , result_prediction)
print("The correct ratio of the mpdel is : " , correct_ratio)
car_feature_names = [ 'O1' , 'D1' , 'O2' , 'D2']
car_target_names = ['0' , '1' ]
dot_data = tree.export_graphviz(clf4, out_file=None, 
                      feature_names = car_feature_names,  
                      class_names=car_target_names,  
                      filled=True, rounded=True,  
                      special_characters=True)  
plot_decision_tree(dot_data , '决策树结果示意图4')
# 保存模型
joblib.dump(clf4,  "DecisionTree4.pkl")   
# lr是训练好的模型， "./ML/test.pkl"是模型要保存的路径及保存模型的文件名，其中，'pkl' 是sklearn中默认的保存格式gai

The correct ratio of the mpdel is :  0.9340540540540541

['DecisionTree4.pkl']

# 得到准确率
accurancy_list.append(accuracy_score(data_test_target, result_prediction))
precision_list.append(precision_score(data_test_target, result_prediction, average='macro'))
recall_list.append(recall_score(data_test_target, result_prediction, average='macro'))
print('accurancy:' ,accuracy_score(data_test_target, result_prediction, normalize=True))
print('macre-precision:' ,precision_score(data_test_target, result_prediction, average='macro')) # precision
print('macro-recall:' , recall_score(data_test_target,result_prediction, average='macro'))  # recall

accurancy: 0.9340540540540541
macre-precision: 0.8905925925925926
macro-recall: 0.9404620298842586

# 3个特征训练决策树模型
data_train_tmp = data_train[: , [3,4,5]]
data_test_tmp = data_test[: ,[3,4,5]]
#网格搜索法加交叉验证获取最优决策树模型参数
params_grid = {'max_depth':[2,3,4],'min_samples_leaf':[2,5,10,20,40] , 'min_samples_split':[20,50,80,100],\
               'class_weight' : ['balanced'] , 'random_state':[0] }
estimator = tree.DecisionTreeClassifier()
grid_search = GridSearchCV(estimator,  param_grid =params_grid, n_jobs = -1, verbose = 2 ,cv=5)  
grid_search.fit(data_train_tmp , data_train_target)
print(grid_search.best_score_)
print(grid_search.best_params_)
print(grid_search.best_estimator_)

Fitting 5 folds for each of 60 candidates, totalling 300 fits

[Parallel(n_jobs=-1)]: Using backend LokyBackend with 8 concurrent workers.
[Parallel(n_jobs=-1)]: Done  34 tasks      | elapsed:    0.3s

0.9167125041143986
{'class_weight': 'balanced', 'max_depth': 4, 'min_samples_leaf': 20, 'min_samples_split': 80, 'random_state': 0}
DecisionTreeClassifier(class_weight='balanced', max_depth=4,
                       min_samples_leaf=20, min_samples_split=80,
                       random_state=0)

[Parallel(n_jobs=-1)]: Done 300 out of 300 | elapsed:    1.0s finished

clf3 = tree.DecisionTreeClassifier(class_weight='balanced', max_depth=4,
                       min_samples_leaf=20, min_samples_split=80,
                       random_state=0)
clf3 = clf3.fit(data_train_tmp , data_train_target)
result_prediction = clf3.predict(data_test_tmp)
correct_ratio = get_correct_ratio(data_test_target , result_prediction)
print("The correct ratio of the mpdel is : " , correct_ratio)
car_feature_names = [ 'D1' , 'O2' , 'D2']
car_target_names = ['0' , '1']
dot_data = tree.export_graphviz(clf3, out_file=None, 
                      feature_names = car_feature_names,  
                      class_names=car_target_names,  
                      filled=True, rounded=True,  
                      special_characters=True)  
plot_decision_tree(dot_data , '决策树结果示意图3')
# 保存模型
joblib.dump(clf3,  "DecisionTree3.pkl")   
# lr是训练好的模型， "./ML/test.pkl"是模型要保存的路径及保存模型的文件名，其中，'pkl' 是sklearn中默认的保存格式gai

The correct ratio of the mpdel is :  0.9221621621621622

['DecisionTree3.pkl']

# 得到准确率
accurancy_list.append(accuracy_score(data_test_target, result_prediction))
precision_list.append(precision_score(data_test_target, result_prediction, average='macro'))
recall_list.append(recall_score(data_test_target, result_prediction, average='macro'))
print('accurancy:' ,accuracy_score(data_test_target, result_prediction, normalize=True))
print('macre-precision:' ,precision_score(data_test_target, result_prediction, average='macro')) # precision
print('macro-recall:' , recall_score(data_test_target,result_prediction, average='macro'))  # recall

accurancy: 0.9221621621621622
macre-precision: 0.8734305538005582
macro-recall: 0.9344744861137099

# 2个特征训练决策树模型
data_train_tmp = data_train[: , [3,5]]
data_test_tmp = data_test[: , [3,5]]
#网格搜索法加交叉验证获取最优决策树模型参数
params_grid = {'max_depth':[2,3,4],'min_samples_leaf':[2,5,10,20,40] , 'min_samples_split':[20,50,80,100],\
               'class_weight' : ['balanced'], 'random_state':[0] }
estimator = tree.DecisionTreeClassifier()
grid_search = GridSearchCV(estimator,  param_grid =params_grid, n_jobs = -1, verbose = 2 ,cv=5)  
grid_search.fit(data_train_tmp , data_train_target)
print(grid_search.best_score_)
print(grid_search.best_params_)
print(grid_search.best_estimator_)

Fitting 5 folds for each of 60 candidates, totalling 300 fits

[Parallel(n_jobs=-1)]: Using backend LokyBackend with 8 concurrent workers.
[Parallel(n_jobs=-1)]: Done  34 tasks      | elapsed:    0.2s

0.9142767801631131
{'class_weight': 'balanced', 'max_depth': 4, 'min_samples_leaf': 2, 'min_samples_split': 20, 'random_state': 0}
DecisionTreeClassifier(class_weight='balanced', max_depth=4, min_samples_leaf=2,
                       min_samples_split=20, random_state=0)

[Parallel(n_jobs=-1)]: Done 300 out of 300 | elapsed:    1.1s finished

clf2 = tree.DecisionTreeClassifier(class_weight='balanced', max_depth=3, min_samples_leaf=2,
                       min_samples_split=20, random_state=0)
clf2 = clf2.fit(data_train_tmp , data_train_target)
result_prediction = clf2.predict(data_test_tmp)
correct_ratio = get_correct_ratio(data_test_target , result_prediction)
print("The correct ratio of the mpdel is : " , correct_ratio)
car_feature_names = [  'D1' , 'D2' ]
car_target_names = ['0' , '1' ]
dot_data = tree.export_graphviz(clf2, out_file=None, 
                      feature_names = car_feature_names,  
                      class_names=car_target_names,  
                      filled=True, rounded=True,  
                      special_characters=True)  
plot_decision_tree(dot_data , '决策树结果示意图2')
# 保存模型
joblib.dump(clf2,  "DecisionTree2.pkl")   
# lr是训练好的模型， "./ML/test.pkl"是模型要保存的路径及保存模型的文件名，其中，'pkl' 是sklearn中默认的保存格式gai

The correct ratio of the mpdel is :  0.9210810810810811

['DecisionTree2.pkl']

# 得到准确率
accurancy_list.append(accuracy_score(data_test_target, result_prediction))
precision_list.append(precision_score(data_test_target, result_prediction, average='macro'))
recall_list.append(recall_score(data_test_target, result_prediction, average='macro'))
print('accurancy:' ,accuracy_score(data_test_target, result_prediction, normalize=True))
print('macre-precision:' ,precision_score(data_test_target, result_prediction, average='macro')) # precision
print('macro-recall:' , recall_score(data_test_target,result_prediction, average='macro'))  # recall

accurancy: 0.9210810810810811
macre-precision: 0.8724985540775014
macro-recall: 0.9303881211408409

# 1个特征训练决策树模型
data_train_tmp = data_train[: , [3]]
data_test_tmp = data_test[: , [3]]
#网格搜索法加交叉验证获取最优决策树模型参数
params_grid = {'max_depth':[2,3,4,5,6,7,8,9,10],'min_samples_leaf':[2,5,10,20,40] , 'min_samples_split':[20,50,80,100],\
               'class_weight' : ['balanced'], 'random_state':[0]  }
estimator = tree.DecisionTreeClassifier()
grid_search = GridSearchCV(estimator,  param_grid =params_grid, n_jobs = -1, verbose = 2 ,cv=5)  
grid_search.fit(data_train_tmp , data_train_target)
print(grid_search.best_score_)
print(grid_search.best_params_)
print(grid_search.best_estimator_)

Fitting 5 folds for each of 180 candidates, totalling 900 fits

[Parallel(n_jobs=-1)]: Using backend LokyBackend with 8 concurrent workers.
[Parallel(n_jobs=-1)]: Done  34 tasks      | elapsed:    0.2s

0.8720857257799072
{'class_weight': 'balanced', 'max_depth': 2, 'min_samples_leaf': 2, 'min_samples_split': 20, 'random_state': 0}
DecisionTreeClassifier(class_weight='balanced', max_depth=2, min_samples_leaf=2,
                       min_samples_split=20, random_state=0)

[Parallel(n_jobs=-1)]: Done 882 tasks      | elapsed:    1.9s
[Parallel(n_jobs=-1)]: Done 900 out of 900 | elapsed:    1.9s finished

clf1 = tree.DecisionTreeClassifier(class_weight='balanced', max_depth=2, min_samples_leaf=2,
                       min_samples_split=20, random_state=0)
clf1 = clf1.fit(data_train_tmp , data_train_target)
result_prediction = clf1.predict(data_test_tmp)
correct_ratio = get_correct_ratio(data_test_target , result_prediction)
print("The correct ratio of the mpdel is : " , correct_ratio)
car_feature_names = [  'D1' ]
car_target_names = ['0' , '1' ]
dot_data = tree.export_graphviz(clf1, out_file=None, 
                      feature_names = car_feature_names,  
                      class_names=car_target_names,  
                      filled=True, rounded=True,  
                      special_characters=True)  
plot_decision_tree(dot_data , '决策树结果示意图1')
# 保存模型
joblib.dump(clf1,  "DecisionTree1.pkl")   
# lr是训练好的模型， "./ML/test.pkl"是模型要保存的路径及保存模型的文件名，其中，'pkl' 是sklearn中默认的保存格式gai

The correct ratio of the mpdel is :  0.8918918918918919

['DecisionTree1.pkl']

# 得到准确率
accurancy_list.append(accuracy_score(data_test_target, result_prediction))
precision_list.append(precision_score(data_test_target, result_prediction, average='macro'))
recall_list.append(recall_score(data_test_target, result_prediction, average='macro'))
print('accurancy:' ,accuracy_score(data_test_target, result_prediction, normalize=True))
print('macre-precision:' ,precision_score(data_test_target, result_prediction, average='macro')) # precision
print('macro-recall:' , recall_score(data_test_target,result_prediction, average='macro'))  # recall

accurancy: 0.8918918918918919
macre-precision: 0.8361055237505802
macro-recall: 0.9064513111117052

绘制出使用不同数目的特征的3个模型评价指标的变化情况

# 将使用不同特征时的准确性，召回率等指标的折线图绘制出来
x_axis = [int((9-i)) for i in range(9)]
fig = plt.figure(figsize=(9,4)) # 设置图大小 figsize=(6,3)
my_x_ticks = np.arange(1, 10, 1)
plt.xticks(my_x_ticks)   #设置坐标轴刻度
plt.plot(x_axis , accurancy_list , c='red' , label = 'Accuracy')
plt.plot(x_axis , precision_list , c='blue',label = 'Macro-precision')
plt.plot(x_axis , recall_list , c='green' ,label = 'Macro-recall')
plt.legend(loc='best')
plt.xlabel("Number of features used to train the decision tree model")
plt.show()

不同的规则下的准确率和召回率变化情况

通过上述的模型一共可以得到3条通勤规则，通过gini对3条通勤规则进行从小到大排序，然后分别分析仅1条规则时的准确率和召回率，2条规则时的准确率和召回率，依次类推，并绘制出折线图。在此的准确率等和之前决策树的准确率不同.

# 得到测试集
N = len(df_isodata) # total number of sample
train_num = round(0.8 * N)
test_num = N - train_num
df_train = df_isodata[0 : train_num]
df_test = df_isodata[train_num : N]
df_test.reset_index(drop = True , inplace = True)
df_test.head(5)

	f1	f2	f3	CAR_NUM	DBSCAN	F1	F2	O1	D1	O2	D2	FS	STD_FD	AVG_FD	FastSearch	ISODATA
0	-0.529897	-1.191786	-1.522080	浙A171LP	2	0.142857	0.285714	0.190476	0.190476	0.047619	0.047619	-2.818446	0.577350	1.666667	0	3
1	-0.552349	0.246508	-0.817364	浙AC5H25	0	0.095238	0.523810	0.190476	0.095238	0.047619	0.095238	-2.022806	2.225395	5.571429	1	3
2	-0.074872	1.004867	0.606384	浙AVJ937	-1	0.333333	0.761905	0.428571	0.142857	0.142857	0.095238	-0.412582	2.619637	7.937500	1	2
3	0.206463	-0.898909	1.392067	浙AQ1T88	-1	0.333333	0.523810	0.523810	0.428571	0.238095	0.190476	0.481541	0.707107	2.500000	1	0
4	0.214456	-1.021856	1.401926	浙AH5V81	-1	0.428571	0.809524	0.238095	0.142857	0.523810	0.142857	0.675609	0.000000	2.000000	1	0

# 获得正确的标签
y_true = []
tmp = list(df_test['ISODATA'])
for i in range(len(tmp)):
    if tmp[i] == 1:
        y_true.append(1)
    else:
        y_true.append(0)
y_true

[0,
 0,
 0,
 0,
 0,
 1,
 0,
 0,
 0,
 0,
 0,
 1,
 0,
 1,
 1,
 1,
 0,
 0,
 0,
 0,
 0,
 0,
 0,
 1,
 0,
 0,
 1,
 1,
 0,
 0,
 0,
 0,
 0,
 0,
 0,
 1,
 1,
 0,
 0,
 0,
 1,
 0,
 0,
 0,
 0,
 0,
 1,
 0,
 1,
 0,
 0,
 1,
 0,
 0,
 0,
 0,
 0,
 0,
 0,
 1,
 0,
 1,
 0,
 0,
 0,
 0,
 1,
 1,
 1,
 0,
 0,
 0,
 1,
 0,
 1,
 0,
 0,
 0,
 0,
 0,
 0,
 0,
 0,
 0,
 0,
 0,
 1,
 0,
 0,
 1,
 0,
 0,
 0,
 0,
 1,
 0,
 0,
 1,
 0,
 1,
 0,
 0,
 0,
 0,
 0,
 0,
 1,
 0,
 0,
 1,
 0,
 0,
 0,
 0,
 0,
 0,
 0,
 0,
 0,
 0,
 1,
 1,
 0,
 0,
 0,
 0,
 0,
 0,
 0,
 1,
 0,
 0,
 0,
 0,
 0,
 0,
 1,
 1,
 1,
 0,
 0,
 0,
 0,
 1,
 0,
 0,
 0,
 0,
 1,
 0,
 0,
 1,
 0,
 0,
 1,
 0,
 0,
 1,
 0,
 0,
 0,
 1,
 0,
 0,
 1,
 1,
 0,
 0,
 0,
 0,
 0,
 0,
 0,
 0,
 0,
 0,
 0,
 1,
 0,
 1,
 1,
 0,
 0,
 0,
 0,
 1,
 1,
 0,
 0,
 0,
 0,
 0,
 0,
 0,
 0,
 0,
 1,
 0,
 0,
 0,
 0,
 0,
 0,
 0,
 0,
 0,
 1,
 0,
 1,
 1,
 0,
 0,
 0,
 0,
 0,
 1,
 0,
 0,
 1,
 0,
 0,
 1,
 0,
 0,
 0,
 0,
 1,
 0,
 0,
 1,
 0,
 0,
 0,
 0,
 0,
 0,
 0,
 0,
 0,
 0,
 0,
 0,
 0,
 0,
 0,
 0,
 0,
 0,
 0,
 0,
 0,
 0,
 0,
 0,
 0,
 0,
 0,
 0,
 0,
 0,
 0,
 0,
 0,
 1,
 0,
 0,
 0,
 0,
 0,
 0,
 0,
 0,
 0,
 0,
 0,
 0,
 0,
 1,
 0,
 0,
 0,
 0,
 0,
 0,
 0,
 0,
 0,
 1,
 0,
 0,
 1,
 0,
 0,
 1,
 0,
 0,
 0,
 0,
 0,
 0,
 0,
 0,
 0,
 0,
 1,
 0,
 0,
 1,
 1,
 0,
 0,
 1,
 0,
 0,
 0,
 0,
 1,
 1,
 1,
 0,
 0,
 0,
 0,
 0,
 0,
 0,
 0,
 1,
 1,
 0,
 1,
 0,
 0,
 0,
 0,
 0,
 0,
 0,
 0,
 0,
 0,
 0,
 1,
 0,
 0,
 0,
 0,
 0,
 1,
 0,
 1,
 0,
 0,
 1,
 0,
 0,
 0,
 0,
 0,
 1,
 0,
 1,
 0,
 1,
 0,
 0,
 0,
 0,
 1,
 0,
 1,
 0,
 0,
 1,
 1,
 0,
 1,
 1,
 0,
 0,
 0,
 1,
 1,
 1,
 0,
 0,
 0,
 0,
 1,
 1,
 0,
 0,
 0,
 0,
 1,
 0,
 0,
 0,
 0,
 0,
 0,
 0,
 0,
 0,
 0,
 0,
 1,
 1,
 0,
 1,
 0,
 0,
 0,
 1,
 1,
 0,
 0,
 0,
 0,
 0,
 1,
 0,
 0,
 0,
 0,
 0,
 1,
 0,
 1,
 0,
 0,
 0,
 0,
 0,
 0,
 0,
 0,
 1,
 1,
 1,
 0,
 0,
 0,
 0,
 0,
 0,
 0,
 0,
 0,
 1,
 0,
 0,
 1,
 0,
 0,
 0,
 0,
 0,
 0,
 0,
 0,
 1,
 0,
 0,
 0,
 0,
 0,
 0,
 1,
 0,
 1,
 0,
 1,
 0,
 0,
 1,
 1,
 0,
 0,
 1,
 1,
 0,
 1,
 0,
 0,
 0,
 0,
 1,
 0,
 0,
 0,
 1,
 0,
 0,
 0,
 0,
 0,
 0,
 0,
 0,
 0,
 0,
 0,
 1,
 0,
 0,
 0,
 1,
 1,
 0,
 0,
 0,
 0,
 1,
 0,
 0,
 0,
 0,
 0,
 1,
 0,
 1,
 1,
 0,
 0,
 0,
 0,
 0,
 0,
 0,
 0,
 0,
 0,
 0,
 0,
 0,
 0,
 0,
 0,
 0,
 0,
 0,
 0,
 0,
 0,
 0,
 0,
 1,
 0,
 0,
 0,
 1,
 1,
 1,
 0,
 0,
 1,
 0,
 1,
 0,
 0,
 0,
 0,
 0,
 0,
 0,
 0,
 0,
 0,
 0,
 0,
 0,
 0,
 0,
 0,
 0,
 0,
 0,
 0,
 0,
 0,
 1,
 0,
 0,
 1,
 0,
 0,
 1,
 0,
 0,
 0,
 1,
 1,
 0,
 0,
 1,
 0,
 0,
 1,
 1,
 0,
 0,
 0,
 0,
 0,
 0,
 1,
 1,
 0,
 0,
 0,
 0,
 0,
 0,
 0,
 0,
 0,
 1,
 0,
 0,
 0,
 0,
 0,
 0,
 1,
 0,
 0,
 0,
 1,
 1,
 0,
 0,
 1,
 0,
 0,
 0,
 0,
 0,
 0,
 0,
 0,
 0,
 0,
 0,
 1,
 0,
 0,
 0,
 0,
 1,
 1,
 0,
 1,
 0,
 0,
 0,
 0,
 0,
 0,
 0,
 1,
 1,
 0,
 1,
 0,
 0,
 0,
 0,
 0,
 1,
 0,
 0,
 0,
 0,
 0,
 0,
 1,
 0,
 0,
 0,
 0,
 0,
 1,
 1,
 0,
 0,
 0,
 0,
 1,
 0,
 1,
 0,
 1,
 1,
 1,
 0,
 1,
 1,
 0,
 0,
 0,
 0,
 1,
 0,
 1,
 1,
 0,
 1,
 0,
 1,
 0,
 0,
 0,
 0,
 0,
 1,
 0,
 0,
 0,
 0,
 0,
 0,
 0,
 0,
 0,
 1,
 0,
 0,
 0,
 0,
 1,
 1,
 0,
 1,
 0,
 1,
 0,
 0,
 0,
 0,
 0,
 1,
 0,
 0,
 0,
 1,
 0,
 0,
 0,
 0,
 0,
 0,
 0,
 0,
 0,
 1,
 0,
 0,
 1,
 0,
 0,
 0,
 0,
 0,
 1,
 0,
 0,
 1,
 0,
 0,
 0,
 1,
 0,
 0,
 0,
 0,
 1,
 0,
 0,
 1,
 1,
 0,
 1,
 0,
 1,
 0,
 1,
 0,
 1,
 0,
 0,
 0,
 0,
 0,
 0,
 0,
 1,
 0,
 0,
 1,
 0,
 0,
 1,
 1,
 0,
 0,
 0,
 0,
 0,
 0,
 0,
 0,
 0,
 0,
 0,
 1,
 0,
 0,
 1,
 0,
 1,
 0,
 0,
 0,
 0,
 0,
 0,
 0,
 0,
 0,
 0,
 0,
 1,
 0,
 0,
 0,
 0,
 0,
 0,
 0,
 0,
 0,
 0,
 1,
 0,
 0,
 0,
 0,
 1,
 0,
 0,
 0,
 0,
 0,
 0,
 0,
 0,
 1,
 0,
 0,
 0,
 0,
 1,
 1,
 0,
 1,
 1,
 0,
 0,
 0,
 0,
 0,
 0,
 1,
 0,
 0,
 0,
 0,
 0,
 0,
 0,
 0,
 0,
 0,
 0,
 1,
 0,
 0,
 0,
 0,
 0,
 0,
 0,
 0,
 0,
 0,
 1,
 0,
 0,
 0,
 0,
 1,
 0,
 0,
 0,
 0,
 0,
 1,
 1,
 0,
 1,
 0,
 0,
 0,
 0,
 0,
 0,
 0,
 0,
 0,
 0,
 0,
 1,
 1,
 0]

accurancy_list2 = []
precision_list2 = []
recall_list2 = []

#使用3条规则，即完整的决策树模型进行预测
y_pred = []
for i in range(len(df_test)):
    tmp = df_test.iloc[i,:]
    if tmp['D1'] > 0.31 and tmp['D2'] > 0.214 :  # rule1
        y_pred.append(1)
    elif tmp['D2'] > 0.595 and tmp['D1'] <= 0.31 : # rule2
        y_pred.append(1)
    elif tmp['D1'] > 0.548 and tmp['D2'] <= 0.214: #rule3
        y_pred.append(1)
    else:
        y_pred.append(0)
accurancy_list2.append(accuracy_score(y_true, y_pred))
precision_list2.append(precision_score(y_true, y_pred, average='macro'))
recall_list2.append(recall_score(y_true, y_pred, average='macro'))
print('accurancy:' , accuracy_score(y_true, y_pred))  # accurancy
print('macre-precision:' ,precision_score(y_true, y_pred, average='macro')) # precision
print('macro-recall:' , recall_score(y_true, y_pred, average='macro'))  # recall

accurancy: 0.9210810810810811
macre-precision: 0.8724985540775014
macro-recall: 0.9303881211408409

#使用2条规则，即完整的决策树模型进行预测
y_pred = []
for i in range(len(df_test)):
    tmp = df_test.iloc[i,:]
    if tmp['D1'] > 0.31 and tmp['D2'] > 0.214 :  # rule1
        y_pred.append(1)
    elif tmp['D2'] > 0.595 and tmp['D1'] <= 0.31 : # rule2
        y_pred.append(1)
    else:
        y_pred.append(0)
accurancy_list2.append(accuracy_score(y_true, y_pred))
precision_list2.append(precision_score(y_true, y_pred, average='macro'))
recall_list2.append(recall_score(y_true, y_pred, average='macro'))
print('accurancy:' , accuracy_score(y_true, y_pred))  # accurancy
print('macre-precision:' ,precision_score(y_true, y_pred, average='macro')) # precision
print('macro-recall:' , recall_score(y_true, y_pred, average='macro'))  # recall

accurancy: 0.9232432432432433
macre-precision: 0.8781775835921138
macro-recall: 0.9199266258587047

#使用1条规则，即完整的决策树模型进行预测
y_pred = []
for i in range(len(df_test)):
    tmp = df_test.iloc[i,:]
    if tmp['D1'] > 0.31 and tmp['D2'] > 0.214 :  # rule1
        y_pred.append(1)
    else:
        y_pred.append(0)
accurancy_list2.append(accuracy_score(y_true, y_pred))
precision_list2.append(precision_score(y_true, y_pred, average='macro'))
recall_list2.append(recall_score(y_true, y_pred, average='macro'))
print('accurancy:' , accuracy_score(y_true, y_pred))  # accurancy
print('macre-precision:' ,precision_score(y_true, y_pred, average='macro')) # precision
print('macro-recall:' , recall_score(y_true, y_pred, average='macro'))  # recall

accurancy: 0.9264864864864865
macre-precision: 0.888095238095238
macro-recall: 0.9084694341236534

# 将使用不同条数的rule时的准确性，召回率等指标的折线图绘制出来
x_axis = [int((3-i)) for i in range(3)]
fig = plt.figure(figsize=(6,3)) # 设置图大小 figsize=(6,3)
my_x_ticks = np.arange(0, 4, 1)
plt.xticks(my_x_ticks)   #设置坐标轴刻度
plt.plot(x_axis , accurancy_list2 , c='red' , label = 'Accuracy')
plt.plot(x_axis , precision_list2 , c='blue',label = 'Macro-precision')
plt.plot(x_axis , recall_list2 , c='green' ,label = 'Macro-recall')
plt.legend(loc='best')
plt.xlabel("Number of commuting rules")
plt.show()

得到决策树模型的更多信息

# 划分训练集,测试集
N = len(df_isodata) # total number of sample
train_num = round(0.8 * N)
test_num = N - train_num
df_train = df_isodata[0 : train_num]
df_test = df_isodata[train_num : N]
data_train_target = df_train['ISODATA'].tolist()
data_train_target = [int(i) for i in data_train_target]
for i in range(len(data_train_target)):
    if data_train_target[i] == 1:
        data_train_target[i] = 1
    else:
        data_train_target[i] = 0
data_train = df_train[['F1' , 'F2' , 'O1' , 'D1' , 'O2' , 'D2' , 'FS' , 'STD_FD' , 'AVG_FD']].values
#data_train = data_train.tolist()

data_test_target = df_test['ISODATA'].tolist()
data_test_target = [int(i) for i in data_test_target]
for i in range(len(data_test_target)):
    if data_test_target[i] == 1:
        data_test_target[i] = 1
    else:
        data_test_target[i] = 0
data_test = df_test[['F1' , 'F2' , 'O1' , 'D1' , 'O2' , 'D2' , 'FS' , 'STD_FD' , 'AVG_FD']].values
#data_test = data_test.tolist()

len(df_train[(df_train['D1'] <= 0.31)&(df_train['D2'] <= 0.405)&(df_train['ISODATA'] != 1)])

2378

len(df_train[(df_train['D1'] <= 0.31)&(df_train['D2'] > 0.405)&(df_train['D2']<=0.595)&(df_train['ISODATA'] != 1)])

94

len(df_train[(df_train['D1'] <= 0.31)&(df_train['D2']>0.405)&(df_train['D2']>0.595)&(df_train['ISODATA'] != 1)])

21

len(df_train[(df_train['D1'] > 0.31)&(df_train['D2']<=0.214)&(df_train['D1']<=0.548)&(df_train['ISODATA'] != 1)])

164

len(df_train[(df_train['D1'] > 0.31)&(df_train['D2']<=0.214)&(df_train['D1']>0.548)&(df_train['ISODATA'] != 1)])

55

len(df_train[(df_train['D1'] > 0.31)&(df_train['D2']>0.214)&(df_train['ISODATA'] != 1)])

183